Title: | Mixed-Effect Models, with or without Spatial Random Effects |
---|---|
Description: | Inference based on models with or without spatially-correlated random effects, multivariate responses, or non-Gaussian random effects (e.g., Beta). Variation in residual variance (heteroscedasticity) can itself be represented by a mixed-effect model. Both classical geostatistical models (Rousset and Ferdy 2014 <doi:10.1111/ecog.00566>), and Markov random field models on irregular grids (as considered in the 'INLA' package, <https://www.r-inla.org>), can be fitted, with distinct computational procedures exploiting the sparse matrix representations for the latter case and other autoregressive models. Laplace approximations are used for likelihood or restricted likelihood. Penalized quasi-likelihood and other variants discussed in the h-likelihood literature (Lee and Nelder 2001 <doi:10.1093/biomet/88.4.987>) are also implemented. |
Authors: | François Rousset [aut, cre, cph] , Jean-Baptiste Ferdy [aut, cph], Alexandre Courtiol [aut] |
Maintainer: | François Rousset <[email protected]> |
License: | CeCILL-2 |
Version: | 4.5.0 |
Built: | 2024-11-07 03:26:48 UTC |
Source: | https://github.com/cran/spaMM |
This is used in tests/test-adjacency-long.R
data("adjlg")
data("adjlg")
Includes an adjacency matrix adjlgMat
. and a data frame adjlg
with 5474 observations on the following 8 variables.
ID
a factor with levels 1
to 1000
months
a numeric vector
GENDER
a character vector
AGE
a numeric vector
X1
a numeric vector
X2
a numeric vector
month
a numeric vector
BUY
a numeric vector
The simulation code shown below is derived from an example produced by Jeroen van den Ochtend. Following a change incorporated in spaMM version 3.8.0, that implied stricter checks of the input matrix, it appeared that the precision matrix generated by this example had inappropriate (repeated) dimnames. This example was then updated to reproduce past fitting results with a correctly formatted matrix. Note that changing the names of an adjacency matrix (as below) is generally unwise as it generally changes the statistical model because these names are matched whenever possible to levels of the grouping factor in the data.
The code was also modified to compensate for changes in R's default random number generator.
data(adjlg) ## See further usage in tests/test-adjacency-long.R ## Not run: # as produced by: library(data.table) ## Included data produced using version 1.10.4.3 library(igraph) ## Included data produced using version 1.2.1 rsample <- function(N=100, ## size of implied adjacency matrix month_max=10,seed) { if (is.integer(seed)) set.seed(seed) dt <- data.table(ID=factor(1:N)) dt$months <- sample(1:month_max,N,replace=T) ## # of liens for each level of ID dt$GENDER <- sample(c("MALE","FEMALE"),N,replace=TRUE) dt$AGE <- sample(18:99,N,replace=T) dt$X1 <- sample(1000:9900,N,replace=T) dt$X2 <- runif(N) dt <- dt[, c(.SD, month=data.table(seq(from=1, to=months, by = 1))), by = ID] dt[,BUY := 0] dt[month.V1==months,BUY := sample(c(0,1),1),by=ID] setnames(dt,"month.V1","month") #### create adjacency matrix Network <- data.table(OUT=sample(dt$ID,N*month_max*4/10)) Network$IN <- sample(dt$ID,N*month_max*4/10) Network <- Network[IN != OUT] Network <- unique(Network) g <- graph.data.frame(Network,directed=F) g <- add_vertices(g,sum(!unique(dt$ID) %in% V(g)), name=unique(dt[!dt$ID %in% V(g),list(ID)])) # => improper names Network <- as_adjacency_matrix(g,sparse = TRUE,type="both") colnames(Network) <- rownames(Network) <- seq(nrow(Network)) # post-v3.8.0 names return(list(data=dt,adjMatrix=Network)) } RNGkind("Mersenne-Twister", "Inversion", "Rounding" ) set.seed(123) adjlg_sam <- rsample(N=1000,seed=NULL) RNGkind("Mersenne-Twister", "Inversion", "Rejection" ) # adjlg <- as.data.frame(adjlg_sam$data) adjlgMat <- adjlg_sam$adjMatrix ## End(Not run)
data(adjlg) ## See further usage in tests/test-adjacency-long.R ## Not run: # as produced by: library(data.table) ## Included data produced using version 1.10.4.3 library(igraph) ## Included data produced using version 1.2.1 rsample <- function(N=100, ## size of implied adjacency matrix month_max=10,seed) { if (is.integer(seed)) set.seed(seed) dt <- data.table(ID=factor(1:N)) dt$months <- sample(1:month_max,N,replace=T) ## # of liens for each level of ID dt$GENDER <- sample(c("MALE","FEMALE"),N,replace=TRUE) dt$AGE <- sample(18:99,N,replace=T) dt$X1 <- sample(1000:9900,N,replace=T) dt$X2 <- runif(N) dt <- dt[, c(.SD, month=data.table(seq(from=1, to=months, by = 1))), by = ID] dt[,BUY := 0] dt[month.V1==months,BUY := sample(c(0,1),1),by=ID] setnames(dt,"month.V1","month") #### create adjacency matrix Network <- data.table(OUT=sample(dt$ID,N*month_max*4/10)) Network$IN <- sample(dt$ID,N*month_max*4/10) Network <- Network[IN != OUT] Network <- unique(Network) g <- graph.data.frame(Network,directed=F) g <- add_vertices(g,sum(!unique(dt$ID) %in% V(g)), name=unique(dt[!dt$ID %in% V(g),list(ID)])) # => improper names Network <- as_adjacency_matrix(g,sparse = TRUE,type="both") colnames(Network) <- rownames(Network) <- seq(nrow(Network)) # post-v3.8.0 names return(list(data=dt,adjMatrix=Network)) } RNGkind("Mersenne-Twister", "Inversion", "Rounding" ) set.seed(123) adjlg_sam <- rsample(N=1000,seed=NULL) RNGkind("Mersenne-Twister", "Inversion", "Rejection" ) # adjlg <- as.data.frame(adjlg_sam$data) adjlgMat <- adjlg_sam$adjMatrix ## End(Not run)
get_any_IC
computes model selection/information criteria such as AIC. See Details for more information about these criteria. The other extractors AIC
and extractAIC
are methods for HLfit
objects of generic functions defined in other packages: AIC
is equivalent to get_any_IC
(for a single fitted-model object), and extractAIC
returns the marginal AIC and the number of degrees of freedom for the fixed effects.
get_any_IC(object, nsim=0L, ..., verbose=interactive(), also_cAIC=TRUE, short.names=NULL) ## S3 method for class 'HLfit' AIC(object, ..., nsim=0L, k, verbose=interactive(), also_cAIC=TRUE, short.names=NULL) ## S3 method for class 'HLfit' extractAIC(fit, scale, k, ..., verbose=FALSE)
get_any_IC(object, nsim=0L, ..., verbose=interactive(), also_cAIC=TRUE, short.names=NULL) ## S3 method for class 'HLfit' AIC(object, ..., nsim=0L, k, verbose=interactive(), also_cAIC=TRUE, short.names=NULL) ## S3 method for class 'HLfit' extractAIC(fit, scale, k, ..., verbose=FALSE)
object , fit
|
A object of class |
scale , k
|
Currently ignored, but are required in the definitions for consistency with the generic. |
verbose |
Whether to print the model selection criteria or not. |
also_cAIC |
Whether to include the plug-in estimate of conditional AIC in the result (its computation may be slow). |
nsim |
Controls whether to include the bootstrap estimate of conditional AIC (see Details) in the result. If positive, |
short.names |
NULL, or boolean; controls whether the return value uses short names ( |
... |
For |
The AIC is a measure (by Kullback-Leibler directed distance, up to an additive constant) of quality of prediction of new data by a fitted model. Comparing information criteria may be viewed as a fast alternative to a comparison of the predictive accuracy of different models by cross-validation. Further procedures for model choice may also be useful (e.g. Williams, 1970; Lewis et al. 2010).
The conditional AIC (Vaida and Blanchard 2005) applies the AIC concept to new realizations of a mixed model, conditional on the realized values of the random effects. Lee et al. (2006) and Ha et al (2007) defined a corrected AIC [i.e., AIC(D*) in their eq. 7] which is here interpreted as the conditional AIC.
Such Kullback-Leibler relative distances cannot generally be evaluated exactly and various estimates have been discussed.
get_any_IC
computes, optionally prints, and returns invisibly one or more of the following quantities:
* Akaike's classical AIC (marginal AIC, mAIC
, i.e., minus twice the marginal log-likelihood plus twice the number of fitted parameters);
* a plug-in estimate (cAIC
) and/or a bootstrap estimate (b_cAIC
) of the conditional AIC;
* a focussed AIC for dispersion parameters (dispersion AIC, dAIC
).
For the conditional AIC, Vaida and Blanchard's plug-in estimator involves the conditional likelihood, and degrees of freedom for (i) estimated residual error parameters and (ii) the overall linear predictor characterized by the Effective degrees of freedom already discussed by previous authors including Lee and Nelder (1996), which gave a plug-in estimator () for it in HGLMs.
By default, the plug-in estimate of both the conditional AIC and of
(
GoFdf
, where is the length of the response vector) are returned by
get_any_IC
. But these are biased estimates of conditional AIC and effective df, and an alternative procedure is available for GLM response families if a non-default positive nsim
value is used. In that case, the conditional AIC is estimated by a bootstrap version of Saefken et al. (2014)'s equation 2.5; this involves refitting the model to each bootstrap samples, so it may take time, and a full cross-validation procedure might as well be considered for model selection.
The dispersion AIC has been defined from restricted likelihood by Ha et al (2007; eq.10). The present implementation will use restricted likelihood only if made available by an REML fit, otherwise marginal likelihood is used.
get_any_IC
, a numeric vector whose possible elements are described in the Details, and whose names are controlled by the short.names
argument. Note that the bootstrap computation actually makes sense and works also for fixed-effect models (although it is not clear how useful it is in that case). The return value will still refer to its results as conditional AIC.
For AIC
, If just one fit object is provided, the same return value as for get_any_IC
. If multiple objects are provided, a data.frame built from such vectors, with rows corresponding to the objects.
For extractAIC
, a numeric vector of length 2, with first and second elements giving
* edf |
the degree of freedom of the fixed-effect terms of the model
for the fitted model |
* AIC |
the (marginal) Akaike Information Criterion for |
Likelihood is broadly defined up to a constant, which opens the way for inconsistency between different likelihood and AIC computations. In spaMM, likelihood is nothing else than the probability or probability density of the data as function of model parameters. No constant is ever added, in contrast to stats::extractAIC
output, so there are discrepancies with the latter function (see Examples).
Ha, I. D., Lee, Y. and MacKenzie, G. (2007) Model selection for multi-component frailty models. Statistics in Medicine 26: 4790-4807.
Lee Y. and Nelder. J. A. 1996. Hierarchical generalized linear models (with discussion). J. R. Statist. Soc. B, 58: 619-678.
Lewis, F., Butler, A. and Gilbert, L. (2011), A unified approach to model selection using the likelihood ratio test. Methods in Ecology and Evolution, 2: 155-162. doi:10.1111/j.2041-210X.2010.00063.x
Saefken B., Kneib T., van Waveren C.-S., Greven S. (2014) A unifying approach to the estimation of the conditional Akaike information in generalized linear mixed models. Electron. J. Statist. 8, 201-225.
Vaida, F., and Blanchard, S. (2005) Conditional Akaike information for mixed-effects models. Biometrika 92, 351-370.
Williams D.A. (1970) Discrimination between regression models to determine the pattern of enzyme synthesis in synchronous cell cultures. Biometrics 26: 23-32.
data("wafers") m1 <- fitme(y ~ X1+X2+X3+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log)) get_any_IC(m1) # => The plug-in estimate is stored in the 'm1' object # as a result of the previous computation, and is now returned even by: get_any_IC(m1, also_cAIC=FALSE) if (spaMM.getOption("example_maxtime")>4) { get_any_IC(m1, nsim=100L, seed=123) # provides bootstrap estimate of cAIC. # (parallelisation options could be used, e.g. nb_cores=detectCores(logical=FALSE)-1L) } extractAIC(m1) ## Not run: # Checking (in)consistency with glm example from help("stats::extractAIC"): utils::example(glm) # => provides 'glm.D93' fit object logLik(glm.D93) # logL= -23.38066 (df=5) dataf <- data.frame(counts=counts,outcome=outcome, treatment=treatment) extractAIC(fitme(counts ~ outcome + treatment, family = poisson(), data=dataf)) # => 56.76132 = -2 logL + 2* df extractAIC(glm.D93) # 56.76132 too # # But for LM: lm.D93 <- lm(counts ~ outcome + treatment, data=dataf) logLik(lm.D93) # logL=-22.78576 (df=6) extractAIC(fitme(counts ~ outcome + treatment, data=dataf)) # 57.5715 = -2 logL + 2* df extractAIC(lm.D93) # 30.03062 ### Inconsistency also apparent in drop1 output for : # Toy data from McCullagh & Nelder (1989, pp. 300-2), as in 'glm' doc: clotting <- data.frame( u = c(5,10,15,20,30,40,60,80,100), lot1 = c(118,58,42,35,27,25,21,19,18), lot2 = c(69,35,26,21,18,16,13,12,12)) # drop1( fitme(lot1 ~ log(u), data = clotting), test = "F") # agains reports marginal AIC # => this may differ strongly from those returned by drop1( < glm() fit > ), # but the latter are not even consistent with those from drop1( < lm() fit > ) # for linear models. Compare drop1( lm(lot1 ~ log(u), data = clotting), test = "F") # consistent with drop1.HLfit() drop1( glm(lot1 ~ log(u), data = clotting), test = "F") # inconsistent ## Discrepancies in drop1 output with Gamma() family: gglm <- glm(lot1 ~ 1, data = clotting, family=Gamma()) logLik(gglm) # -40.34633 (df=2) spgglm <- fitme(lot1 ~ 1, data = clotting, family=Gamma()) logLik(spgglm) # -40.33777 (slight difference: # see help("method") for difference in estimation method between glm() and fitme()). # Yet this does not explain the following: drop1( fitme(lot1 ~ log(u), data = clotting, family=Gamma()), test = "F") # => second AIC is 84.676 as expected from above logLik(spgglm). drop1( glm(lot1 ~ log(u), data = clotting, family=Gamma()), test = "F") # => second AIC is 1465.27, quite different from -2*logLik(gglm) + 2*df ## End(Not run)
data("wafers") m1 <- fitme(y ~ X1+X2+X3+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log)) get_any_IC(m1) # => The plug-in estimate is stored in the 'm1' object # as a result of the previous computation, and is now returned even by: get_any_IC(m1, also_cAIC=FALSE) if (spaMM.getOption("example_maxtime")>4) { get_any_IC(m1, nsim=100L, seed=123) # provides bootstrap estimate of cAIC. # (parallelisation options could be used, e.g. nb_cores=detectCores(logical=FALSE)-1L) } extractAIC(m1) ## Not run: # Checking (in)consistency with glm example from help("stats::extractAIC"): utils::example(glm) # => provides 'glm.D93' fit object logLik(glm.D93) # logL= -23.38066 (df=5) dataf <- data.frame(counts=counts,outcome=outcome, treatment=treatment) extractAIC(fitme(counts ~ outcome + treatment, family = poisson(), data=dataf)) # => 56.76132 = -2 logL + 2* df extractAIC(glm.D93) # 56.76132 too # # But for LM: lm.D93 <- lm(counts ~ outcome + treatment, data=dataf) logLik(lm.D93) # logL=-22.78576 (df=6) extractAIC(fitme(counts ~ outcome + treatment, data=dataf)) # 57.5715 = -2 logL + 2* df extractAIC(lm.D93) # 30.03062 ### Inconsistency also apparent in drop1 output for : # Toy data from McCullagh & Nelder (1989, pp. 300-2), as in 'glm' doc: clotting <- data.frame( u = c(5,10,15,20,30,40,60,80,100), lot1 = c(118,58,42,35,27,25,21,19,18), lot2 = c(69,35,26,21,18,16,13,12,12)) # drop1( fitme(lot1 ~ log(u), data = clotting), test = "F") # agains reports marginal AIC # => this may differ strongly from those returned by drop1( < glm() fit > ), # but the latter are not even consistent with those from drop1( < lm() fit > ) # for linear models. Compare drop1( lm(lot1 ~ log(u), data = clotting), test = "F") # consistent with drop1.HLfit() drop1( glm(lot1 ~ log(u), data = clotting), test = "F") # inconsistent ## Discrepancies in drop1 output with Gamma() family: gglm <- glm(lot1 ~ 1, data = clotting, family=Gamma()) logLik(gglm) # -40.34633 (df=2) spgglm <- fitme(lot1 ~ 1, data = clotting, family=Gamma()) logLik(spgglm) # -40.33777 (slight difference: # see help("method") for difference in estimation method between glm() and fitme()). # Yet this does not explain the following: drop1( fitme(lot1 ~ log(u), data = clotting, family=Gamma()), test = "F") # => second AIC is 84.676 as expected from above logLik(spgglm). drop1( glm(lot1 ~ log(u), data = clotting, family=Gamma()), test = "F") # => second AIC is 1465.27, quite different from -2*logLik(gglm) + 2*df ## End(Not run)
Autocorrelated gaussian random effects can be specified in terms of their covariance matrix, or in terms of the precision matrix (i.e. inverse covariance matrix). In a pre-processing step, spaMM may assess whether such precision matrices are sparse but the correlation matrix is dense, and if so, it may use “sparse-precision” algorithms efficient for this case. If the precision matrix does not appear sufficiently sparser than the correlation matrix, correlation matrices are used, and they can themselves be sparse or dense, with distinct algebraic methods used in each case.
For example, when the model includes a corrMatrix term specified by a covariance matrix, the precision matrix may be computed to assess its sparseness. The Example below illustrates a case where detecting sparsity of the precision matrix allows a faster fit. However, such a comparison of correlation and precision matrices takes time and is not performed for all types of random-effect structures. Instead, some fast heuristics may be used (see Details). The default selection of methods may not always be optimal, and may be overcome by using the control.HLfit
argument of the fitting function (or by spaMM.options()
, see Details). In particular one can use either control.HLfit=list(sparse_precision= <TRUE|FALSE>)
or
control.HLfit=list(algebra= <"spprec"|"spcorr"|"decorr">)
with the obvious expected effects.
Such control may be useful when you already know that the precision matrix is sparse (as spaMM may even kindly remind you of, see Example below). In that case, it is also efficient to specify the precision matrix directly (see Example in Gryphon
), as spaMM then assumes that sparse-precision methods are better without checking the correlation matrix.
Such control may also be useful when the correlation matrix is nearly singular so that computation of its inverse fails. This may occur if the model is poorly specified, but also occurs sometimes for valid correlation models because inversion of large matrices though Cholesky methods is not numerically accurate enough. In the latter case, you may be directed to this documentation by an error message, and specifying sparse_precision= FALSE
may be useful.
Currently the sparse-precision methods are selected by default in two cases (with possible exceptions indicated by specific messages): (1) for models including IMRF
random effects; and (2) when the corrMatrix
(or covStruct
) syntax is used to provide a fixed precision matrix. Further, for models including autoregressive terms other than IMRF (i.e., adjacency, AR1), sparse-precision methods may or may not be selected on a simple heuristic based on the likely structure of the correlation matrix.
Algebraic methods can be controlled globally over all further fits by usingspaMM.options(sparse_precision= <TRUE|FALSE>)
and, among the correlation-based methods,spaMM.options(QRmethod= <"sparse"|"dense">)
to select "spcorr"
vs. "decorr"
methods. Fit-specific controls (by control.HLfit
) override these global ones.
if (spaMM.getOption("example_maxtime")>6) { data("Gryphon") gry_df <- fitme(BWT ~ 1 + corrMatrix(1|ID), corrMatrix = Gryphon_A, data = Gryphon_df, method = "REML") how(gry_df) # => Note the message about 'Choosing matrix methods...'. # Using control.HLfit=list(algebra="spprec") would indeed # save the time used to select this method. # Conversely, using a correlation-based method would be a waste of time: gry_dn <- fitme(BWT ~ 1 + corrMatrix(1|ID), corrMatrix = Gryphon_A, data = Gryphon_df, method = "REML", control.HLfit=list(sparse_precision=FALSE)) how(gry_dn) # forced dense-correlation methods, which is slower here. }
if (spaMM.getOption("example_maxtime")>6) { data("Gryphon") gry_df <- fitme(BWT ~ 1 + corrMatrix(1|ID), corrMatrix = Gryphon_A, data = Gryphon_df, method = "REML") how(gry_df) # => Note the message about 'Choosing matrix methods...'. # Using control.HLfit=list(algebra="spprec") would indeed # save the time used to select this method. # Conversely, using a correlation-based method would be a waste of time: gry_dn <- fitme(BWT ~ 1 + corrMatrix(1|ID), corrMatrix = Gryphon_A, data = Gryphon_df, method = "REML", control.HLfit=list(sparse_precision=FALSE)) how(gry_dn) # forced dense-correlation methods, which is slower here. }
For 948 “accessions” from European Arabidopsis thaliana populations, this data set merges the genotypic information at four single nucleotide polymorphisms (SNP) putatively involved in adaptation to climate (Fournier-Level et al, 2011, Table 1), with 13 climatic variables from Hancock et al. (2011).
data("arabidopsis")
data("arabidopsis")
The data frame includes 948 observations on the following variables:
Genotypes at four SNP loci
latitude
longitude
Thirteen climatic variables. See Hancock et al. (2011) for details about these variables.
The response is binary so method="PQL/L"
seems warranted (see Rousset and Ferdy, 2014).
The data were retrieved from http://bergelson.uchicago.edu/regmap-data/climate-genome-scan
on 22 February 2013 (they may no longer be available from there).
Fournier-Level A, Korte A., Cooper M. D., Nordborg M., Schmitt J., Wilczek AM (2011). A map of local adaptation in Arabidopsis thaliana. Science 334: 86-89.
Hancock, A. M., Brachi, B., Faure, N., Horton, M. W., Jarymowycz, L. B., Sperone, F. G., Toomajian, C., Roux, F., and Bergelson, J. 2011. Adaptation to climate across the Arabidopsis thaliana genome, Science 334: 83-86.
Rousset F., Ferdy, J.-B. (2014) Testing environmental and genetic effects in the presence of spatial autocorrelation. Ecography, 37: 781-790. doi:10.1111/ecog.00566
data("arabidopsis") if (spaMM.getOption("example_maxtime")>2.5) { fitme(cbind(pos1046738,1-pos1046738)~seasonal+Matern(1|LAT+LONG), fixed=list(rho=0.119278,nu=0.236990,lambda=8.599), family=binomial(),method="PQL/L",data=arabidopsis) } ## The above 'fixed' values are deduced from the following fit: if (spaMM.getOption("example_maxtime")>46) { SNPfit <- fitme(cbind(pos1046738,1-pos1046738)~seasonal+Matern(1|LAT+LONG), verbose=c(TRACE=TRUE), family=binomial(),method="PQL/L",data=arabidopsis) summary(SNPfit) # p_v=-125.0392 }
data("arabidopsis") if (spaMM.getOption("example_maxtime")>2.5) { fitme(cbind(pos1046738,1-pos1046738)~seasonal+Matern(1|LAT+LONG), fixed=list(rho=0.119278,nu=0.236990,lambda=8.599), family=binomial(),method="PQL/L",data=arabidopsis) } ## The above 'fixed' values are deduced from the following fit: if (spaMM.getOption("example_maxtime")>46) { SNPfit <- fitme(cbind(pos1046738,1-pos1046738)~seasonal+Matern(1|LAT+LONG), verbose=c(TRACE=TRUE), family=binomial(),method="PQL/L",data=arabidopsis) summary(SNPfit) # p_v=-125.0392 }
These times-series correlation models can be declared as correlation models for random effect.
The AR(p) model is here parametrized by the partial correlation coefficients of the levels of the random effect, {U_t}, corr(U_s,U_t|U_(s+1),...,U_(t-1)), with valid values in the hypercube ]-1,,1[^p (Barndorff-Nielsen and Schou, 1973).
In the autoregressive-moving average ARMA(p,q) model, the AR part is parametrized in the same way. AR parameters are named "p1"
, "p2"
..., and MA parameters are named "q1"
, "q2"
... .
Implementation of the AR(p) model uses the sparsity of the inverse covariance matrix. In the ARMA(p,q) model, neither the covariance nor its inverse are sparse, so fits are expected to be more time- and memory-consuming.
# corrFamily constructors: ARp(p=1L, fixed=NULL, corr=TRUE, tpar=1/(1+seq(p))) ARMA(p=1L, q=1L, fixed=NULL, tpar=c(1/(1+seq_len(p)),1/(1+seq_len(q))))
# corrFamily constructors: ARp(p=1L, fixed=NULL, corr=TRUE, tpar=1/(1+seq(p))) ARMA(p=1L, q=1L, fixed=NULL, tpar=c(1/(1+seq_len(p)),1/(1+seq_len(q))))
p |
Integer: order of the autoregressive process. |
q |
Integer: order of the moving-average process. |
tpar |
Numeric vector: template values of the partial coefficient coefficients of the autoregressive process, and the traditional coefficients of the moving-average processe, in this order. The |
fixed |
NULL or numeric vector, to fix the parameters of this model. |
corr |
For development purposes, better ignored in normal use. |
The ARp
and ARMA
functions return a corrFamily
descriptor, hence a list
including element Cf
, a function returning, for given ARMA or AR parameters, the correlation matrix for ARMA
, or its inverse for ARp
.
The fitted correlation matrix can be extracted from a fit object, as for any autocorrelated random effect, by Corr(
<fit object>)[[
<random-effect index>]]
.
Barndorff-Nielsen 0. and Schou G., 1973 On the parametrization of autoregressive models by partial autocorrelations. J. Multivariate Analysis 3: 408-419. doi:10.1016/0047-259X(73)90030-4
if (spaMM.getOption("example_maxtime")>2) { ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from 'stats' package ## Default 'tpar' => AR1 model # (ARpfit <- fitme(lh ~ 1 + ARp(1|time), data=ts, method="REML")) # ## which is equivalent to # (AR1fit <- fitme(lh ~ 1 +AR1(1|time), data=ts, method="REML")) ## AR(3) model # (AR3fit <- fitme(lh ~ 1 + ARp(1|time, p=3), data=ts, method="REML")) ## Same but with fixed 2-lag partial autocorrelation # (AR3fix <- fitme(lh ~ 1 + ARp(1|time, p=3, fixed=c(p2=0)), data=ts, method="REML")) # # The fit should be statistically equivalent to # (AR3_fix <- fitme(lh ~ 1 + ARp(1|time, p=3), data=ts, method="REML", fixed=list(corrPars=list("1"=c(p2=0))))) # # with subtle differences in the structure of the fit objects: # get_ranPars(AR3fix)$corrPars # p2 was not a parameter of the model get_ranPars(AR3_fix)$corrPars # p2 was a fixed parameter of the model # # get_fittefPars() expectedly ignores 'p2' whichever way it was fixed. ## Same as 'AR3fix' but with an additional MA(1) component # (ARMAfit <- fitme(lh ~ 1 + ARMA(1|time, p=3, q=1, fixed=c(p2=0)), data=ts, method="REML")) }
if (spaMM.getOption("example_maxtime")>2) { ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from 'stats' package ## Default 'tpar' => AR1 model # (ARpfit <- fitme(lh ~ 1 + ARp(1|time), data=ts, method="REML")) # ## which is equivalent to # (AR1fit <- fitme(lh ~ 1 +AR1(1|time), data=ts, method="REML")) ## AR(3) model # (AR3fit <- fitme(lh ~ 1 + ARp(1|time, p=3), data=ts, method="REML")) ## Same but with fixed 2-lag partial autocorrelation # (AR3fix <- fitme(lh ~ 1 + ARp(1|time, p=3, fixed=c(p2=0)), data=ts, method="REML")) # # The fit should be statistically equivalent to # (AR3_fix <- fitme(lh ~ 1 + ARp(1|time, p=3), data=ts, method="REML", fixed=list(corrPars=list("1"=c(p2=0))))) # # with subtle differences in the structure of the fit objects: # get_ranPars(AR3fix)$corrPars # p2 was not a parameter of the model get_ranPars(AR3_fix)$corrPars # p2 was a fixed parameter of the model # # get_fittefPars() expectedly ignores 'p2' whichever way it was fixed. ## Same as 'AR3fix' but with an additional MA(1) component # (ARMAfit <- fitme(lh ~ 1 + ARMA(1|time, p=3, q=1, fixed=c(p2=0)), data=ts, method="REML")) }
The lmerTest::contest
function, drop1
and anova
methods implement a number of tests for linear mixed models, e.g. using effective degrees of freeedom based on (a generalization of) Satterthwaite's method. These tests can be performed using spaMM fits through the conversion of the fit object, by the as_LMLT
function, to an ad-hoc format acceptable as input to lmerTest's internal procedures. The separately documented drop1.HLfit
and (optionally) anova.HLfit
methods, when called on a single LMM fit object, perform the conversion by as_LMLT
and call drop1
or anova
methods defined by lmerTest.
Only the tests using lmerTest's default method ddf="Satterthwaite"
are formally supported, as the converted object do not have the required format for the other methods. Only LMMs are handled by lmerTest, and residual-dispersion models are not yet handled by the conversion. However, the conversion extends lmerTest's functionality by handling all random-effect parameters handled by numInfo
, therefore including (e.g.) spatial-correlation parameters not handled by lme4.
as_LMLT(fitobject, nuisance=NULL, verbose=TRUE, transf=TRUE, check_deriv=NULL, ...)
as_LMLT(fitobject, nuisance=NULL, verbose=TRUE, transf=TRUE, check_deriv=NULL, ...)
fitobject |
Object of class |
nuisance |
A list of fitted values of parameters that affect the distribution of the test of fixed effects, in the format of the |
verbose |
boolean: controls printing of the message that shows the |
transf |
boolean: whether to evaluate numerical derivatives on a transformed parameter scale, or not (may affect numerical precision). |
check_deriv |
See same-named argument of |
... |
Other arguments that may be needed by some method (currently ignored). |
The value is returned invisibly. It is an S4 object of class "LMLT"
with slots matching those required in objects of S4 class "lmerModLmerTest"
when used by package lmerTest with ddf="Satterthwaite"
(many additional slots of a formal "lmerModLmerTest"
object are missing). The additional nuisance
slot contains the nuisance
list.
Alexandra Kuznetsova, Per B. Brockhoff and Rune H. B. Christensen (2017) lmerTest Package: Tests in Linear Mixed Effects Models. Journal of Statistical Software, 82(13), 1–26. doi:10.18637/jss.v082.i13
## Reproducing an example from the doc of lmerTest::contest.lmerModLmerTest, # using a spaMM fit as input. ## Not run: data("sleepstudy", package="lme4") ## The fit: spfit <- fitme(Reaction ~ Days + I(Days^2) + (1|Subject) + (0+Days|Subject), sleepstudy, method="REML") ## Conversion: spfit_lmlt <- as_LMLT(spfit) ## Functions from package lmerTest can then be called on this object: lmerTest::contest(spfit_lmlt, L=diag(3)[2:3, ]) # Test of 'Days + I(Days^2)'. # anova(spfit_lmlt, type="1") # : using lmerTest:::anova.lmerModLmerTest() drop1(spfit_lmlt) # : using lmerTest:::drop1.lmerModLmerTest() ## End(Not run)
## Reproducing an example from the doc of lmerTest::contest.lmerModLmerTest, # using a spaMM fit as input. ## Not run: data("sleepstudy", package="lme4") ## The fit: spfit <- fitme(Reaction ~ Days + I(Days^2) + (1|Subject) + (0+Days|Subject), sleepstudy, method="REML") ## Conversion: spfit_lmlt <- as_LMLT(spfit) ## Functions from package lmerTest can then be called on this object: lmerTest::contest(spfit_lmlt, L=diag(3)[2:3, ]) # Test of 'Days + I(Days^2)'. # anova(spfit_lmlt, type="1") # : using lmerTest:::anova.lmerModLmerTest() drop1(spfit_lmlt) # : using lmerTest:::drop1.lmerModLmerTest() ## End(Not run)
Diverse autoregressive (AR) models are implemented in spaMM. This documentation describe the adjacency
model (a conditional AR, i.e., CAR), and the AR1
model for time series. Other documentation deals with more or less distantly related models: ARp
for more general AR(p) and ARMA(p,q) models for time series, and IMRF
and MaternIMRFa
for mesh-based approximations of geostatistical models.
An AR1 random effect is specified as AR1(1|<grouping factor>)
. It describes correlations between realizations of the random effect for (typically) successive time-steps by a correlation , denoted
ARphi
in function calls. Nested AR1 effects can be specified by a nested grouping factor, as in AR1(1|<time index> %in% <nesting factor>)
.
A CAR random effect is specified as adjacency(1|<grouping factor>)
. The correlations among levels of the random effect form a
matrix (I
adjMatrix
, in terms of an
adjMatrix
matrix which must be provided, and of the scalar , denoted
rho
in function calls.
The rows and columns of adjMatrix
must have names matching those of levels of the random effect or else are assumed to match a sequence, from 1 to the number of columns, of values of the geographic location index specifying the spatial random effect. For example, if the model formula
isy ~ adjacency(1|geo.loc)
and <data>$geo.loc
is 2,4,3,1,... the first row/column of the matrix refers to geo.loc
=1, i.e. to the fourth row of the data.
Efficient algorithms for CAR models have been widely discussed in particular in the econometric literature (e.g., LeSage and Pace 2009), but these models are not necessarily recommended for irregular lattices (see Wall, 2004 and Martellosio, 2012 for some insights on the implications of autoregressive models).
In CAR models, the covariance matrix of random effects u can be described as (I
W
where W is the (symmetric) adjacency matrix.
HLCor
uses the spectral decomposition of the adjacency matrix, written as W=VDV' where D is a diagonal matrix of eigenvalues . The covariance of V'u is
(I
D
, which is a diagonal matrix with elements
=
/(1
). Hence
is in the linear predictor form
+
This can be used to fit
and
efficiently. A call to
corrHLfit
with the additional argument
init.HLfit=list(rho=0)
should be equivalent in speed and result to the HLCor
call.
This is fast for small datasets (as in the example below) but more generic maximization algorithms may be preferable for large ones. It is suggested to use fitme
generally unless one has a large number of small data sets to analyze. A call to fitme
or corrHLfit
without that initial value does not use the spectral decomposition. It performs numerical maximization of the likelihood (or restricted likelihood) as function of the correlation parameter . The choice of fitting function may slightly impact the results. The ML fits by
corrHLfit
and HLCor
should be practically equivalent. The REML fits should slightly differ from each other, due to the fact that the REML approximation for GLMMs does not maximize a single likelihood function.
If HLCor
is used, the results are reported as the coefficients (
(Intercept)
) and (
adjd
) of the predictor for , in addition to the resulting values of
and of the common
factor.
Different fits may also differ in using or not algorithms that exploit the sparsity of the precision matrix of the autoregressive random effect. By default, spaMM tends to select sparse-precision algorithms for large datasets and large (i.e. many-level) random effects (details are complex). However, for AR1 models, the dimension of the implied precision matrix is determined by the extreme values of grouping factor (typically interpreted as a time index), as all intermediate values must be considered. Then, the correlation-based algorithms may be more efficient if only a few levels are present in the data, as only a small correlation matrix is required in that case.
LeSage, J., Pace, R.K. (2009) Introduction to Spatial Econometrics. Chapman & Hall/CRC.
Martellosio, F. (2012) The correlation structure of spatial autoregressions, Econometric Theory 28, 1373-1391.
Wall M.M. (2004) A close look at the spatial structure implied by the CAR and SAR models: Journal of Statistical Planning and Inference 121: 311-324.
##### AR1 random effect: ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from stats package fitme(lh ~ 1 +AR1(1|time), data=ts, method="REML") # With fixed parameters: # HLCor(lh ~ 1 +AR1(1|time), data=ts, ranPars=list(ARphi=0.5,lambda=0.25,phi=0.001)) ##### CAR random effect: data("scotlip") # CAR by Laplace with 'outer' estimation of rho if (spaMM.getOption("example_maxtime")>0.8) { fitme(cases ~ I(prop.ag/10)+adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix, family=poisson(), data=scotlip) } # CAR by Laplace with 'inner' estimation of rho HLCor(cases ~ I(prop.ag/10)+adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix, family=poisson(), data=scotlip, method="ML")
##### AR1 random effect: ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from stats package fitme(lh ~ 1 +AR1(1|time), data=ts, method="REML") # With fixed parameters: # HLCor(lh ~ 1 +AR1(1|time), data=ts, ranPars=list(ARphi=0.5,lambda=0.25,phi=0.001)) ##### CAR random effect: data("scotlip") # CAR by Laplace with 'outer' estimation of rho if (spaMM.getOption("example_maxtime")>0.8) { fitme(cases ~ I(prop.ag/10)+adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix, family=poisson(), data=scotlip) } # CAR by Laplace with 'inner' estimation of rho HLCor(cases ~ I(prop.ag/10)+adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix, family=poisson(), data=scotlip, method="ML")
Returns a family
object for beta-response models.
The model described by such a family is characterized by a linear predictor, a link function, and the beta density for the residual variation.
The precision parameter prec
of this family is a positive value such that the variance of the response given its mean is
prec
). prec
is thus the precision parameter of Ferrari & Cribari-Neto (2004) and of the betareg package (Cribari-Neto & Zeileis 2010).
A fixed-effect residual-dispersion model can be fitted, using the resid.model
argument, which is used to specify the form of the logarithm of the precision parameter (see Examples). Thus the variance of the response become exp(<specified linear expression>)
).
beta_resp(prec = stop("beta_resp's 'prec' must be specified"), link = "logit")
beta_resp(prec = stop("beta_resp's 'prec' must be specified"), link = "logit")
prec |
Scalar (or left unspecified): precision parameter of the beta distribution. |
link |
logit, probit, cloglog or cauchit link, specified by any of the available ways for GLM links (name, character string, one-element character vector, or object of class |
Prior weights are meaningful for this family and handled as a factor of the precision parameter (as for GLM families) hence here not as a divisor of the variance (in contrast to GLM families): the variance of the response become prec*<prior weights>
). However, this feature is experimental and may be removed in the future. The fitting function's resid.model
argument may be preferred to obtain the same effect, by specifying an offset(log(<prior weights>))
in its formula (given the log link used in that model). As usual in spaMM, the offset(.) argument should be a vector and any variable necessary for evaluating it should be in the data
.
A list, formally of class c("LLF", "family")
. See LL-family
for details about the structure and usage of such objects.
Cribari-Neto, F., & Zeileis, A. (2010). Beta Regression in R. Journal of Statistical Software, 34(2), 1-24. doi:10.18637/jss.v034.i02
Ferrari SLP, Cribari-Neto F (2004). “Beta Regression for Modelling Rates and Proportions.” Journal of Applied Statistics, 31(7), 799-815.
Further examples in LL-family
.
set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE), x_het=runif(100)) fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat) ## same logL, halved 'prec' when prior weights=2 are used: # fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat, prior.weights=rep(2,100)) ## With model for residual dispersion: # fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat, resid.model= ~ x_het)
set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE), x_het=runif(100)) fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat) ## same logL, halved 'prec' when prior weights=2 are used: # fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat, prior.weights=rep(2,100)) ## With model for residual dispersion: # fitme(y ~1+(1|grp), family=beta_resp(), data= beta_dat, resid.model= ~ x_het)
Returns a family
object for beta-binomial models.
The model described by such a family is characterized by a linear predictor, a link function, and the beta-binomial distribution for residual variation.
The precision parameter prec
of this family is a positive value such that the variance of the beta-distributed latent variable given its mean is
prec
). prec
is thus the same precision parameter as for the beta response family (see beta_resp
. The variance of the beta-binomial sample of size is response is
prec
prec
).
A fixed-effect residual-dispersion model can be fitted, using the resid.model
argument, which is used to specify the form of the logarithm of the precision parameter (see Examples). Thus the variance of the latent beta-distributed variable becomes exp(<specified linear expression>)
).
betabin(prec = stop("betabin's 'prec' must be specified"), link = "logit")
betabin(prec = stop("betabin's 'prec' must be specified"), link = "logit")
prec |
Scalar (or left unspecified): precision parameter of the beta distribution. |
link |
logit, probit, cloglog or cauchit link, specified by any of the available ways for GLM links (name, character string, one-element character vector, or object of class |
Prior weights are meaningful for this family and handled as a factor of the precision parameter of the latent beta-distributed variable: the variance of the latent variable become prec*<prior weights>
). However, this feature is experimental and may be removed in the future. The fitting function's resid.model
argument may be preferred to obtain the same effect, by specifying an offset(log(<prior weights>))
in its formula (given the log link used in that model). As usual in spaMM, the offset(.) argument should be a vector and any variable necessary for evaluating it should be in the data
.
A list, formally of class c("LLF", "family")
. See LL-family
for details about the structure and usage of such objects.
if (requireNamespace("agridat", quietly = TRUE)) { data("crowder.seeds", package = "agridat") fitme(cbind(germ,n-germ) ~ gen+extract+(1|plate), data=crowder.seeds, family=betabin()) } else { data(clinics) fitme(cbind(npos,nneg)~1+(1|clinic), family=betabin(), data=clinics) }
if (requireNamespace("agridat", quietly = TRUE)) { data("crowder.seeds", package = "agridat") fitme(cbind(germ,n-germ) ~ gen+extract+(1|plate), data=crowder.seeds, family=betabin()) } else { data(clinics) fitme(cbind(npos,nneg)~1+(1|clinic), family=betabin(), data=clinics) }
This data set is extracted from a study of genetic polymorphisms potentially associated to migration behaviour in the blackcap (Sylvia atricapilla). Across different populations in Europe and Africa, the average migration behaviour was found to correlate with average allele size (dependent on the number of repeats of a small DNA motif) at the locus ADCYAP1, encoding a neuropeptide. This data set is quite small and ill-suited for separating random-effect variance from residual variance. The likelihood surface for the Matérn model actually has local maxima.
data("blackcap")
data("blackcap")
The data frame includes 14 observations on the following variables:
latitude, indeed.
longitude, indeed.
migration status as determined by Mueller et al, from 0 (resident populations) to 2.5 (long-distance migratory populations)
Mean allele sizes in each population
Numerical index for the populations
Migration status was coded as : pure resident populations as '0', resident populations with some migratory restlessness as '0.5', partial migratory populations as '1', completely migratory populations migrating short-distances as '1.5', intermediate-distance migratory populations as '2' and distinct long-distance migratory populations as '2.5'.
Data from Mueller et al. (2011), including supplementary material now available from doi:10.1098/rspb.2010.2567.
Mueller, J. C., Pulido, F., and Kempenaers, B. 2011. Identification of a gene associated with avian migratory behaviour, Proc. Roy. Soc. (Lond.) B 278, 2848-2856.
## see 'fitme', 'corrHLfit' and 'fixedLRT' for examples involving these data
## see 'fitme', 'corrHLfit' and 'fixedLRT' for examples involving these data
The Cauchy family of correlation functions is useful to describe spatial processes with power-law decrease of correlation at long distance. It is valid for Euclidean distances in spaces of any dimension, and for great-circle distances on spheres of any dimension. It has a scale parameter (rho
, as in the Matérn correlation function), a shape
(or “smoothness”, Gneiting 2013) parameter, and a long
-memory dep
endence (or, more abstractly, “shape”; Gneiting 2013) parameter (Gneiting and Schlater 2004). The present implementation also accepts a Nugget
parameter. The family can be invoked in two ways. First, the CauchyCorr
function evaluates correlations, using distances as input. Second, a term of the form Cauchy(1|
<...>)
in a formula
specifies a random effect with Cauchy correlation function, using coordinates found in a data frame as input. In the latter case, the correlations between realizations of the random effect for any two observations in the data will be the value of the Cauchy function at the scaled distance between coordinates specified in <...>, using “+” as separator (e.g., Cauchy(1|longitude+latitude)
). A syntax of the form Cauchy(1|longitude+latitude %in% grp)
can be used to specify a Cauchy random effect with independent realizations for each level of the grouping variable grp
.
## Default S3 method: CauchyCorr(d, rho=1, shape, longdep, Nugget=NULL) # Cauchy(1|...)
## Default S3 method: CauchyCorr(d, rho=1, shape, longdep, Nugget=NULL) # Cauchy(1|...)
d |
Euclidean or great-circle distance |
rho |
The scaling factor for distance, a real >0. |
shape |
The shape (smoothness) parameter, a real 0<.<=2 for Euclidean distances and 0<.<=1 for great-circle distances. Smoothness increases, and fractal dimension decreases, with increasing shape (the fractal dimension of realizations in spaces of dimension |
longdep |
The long-memory dependence parameter, a real >0. It gives the exponent of the asymptotic decrease of correlation with distance: the smaller |
Nugget |
(Following the jargon of Kriging) a parameter describing a discontinuous decrease in
correlation at zero distance. Correlation will always be 1 at |
... |
Names of coordinates, using “+” as separator (e.g., |
The correlation at distance is
Scalar/vector/matrix depending on input.
Gneiting, T. and Schlater M. (2004) Stochastic models that separate fractal dimension and the Hurst effect. SIAM Rev. 46: 269–282.
Gneiting T. (2013) Strictly and non-strictly positive definite functions on spheres. Bernoulli 19: 1327-1349.
data("blackcap") fitme(migStatus ~ means+ Cauchy(1|longitude+latitude), data=blackcap, # fixed=list(longdep=0.5,shape=0.5,rho=0.05) ) ## The Cauchy family can be used in Euclidean spaces of any dimension: set.seed(123) randpts <- matrix(rnorm(20),nrow=5) distMatrix <- as.matrix(proxy::dist(randpts)) CauchyCorr(distMatrix,rho=0.1,shape=1,longdep=10) # See ?MaternCorr for examples of syntaxes for group-specific random effects, # also handled by Cauchy().
data("blackcap") fitme(migStatus ~ means+ Cauchy(1|longitude+latitude), data=blackcap, # fixed=list(longdep=0.5,shape=0.5,rho=0.05) ) ## The Cauchy family can be used in Euclidean spaces of any dimension: set.seed(123) randpts <- matrix(rnorm(20),nrow=5) distMatrix <- as.matrix(proxy::dist(randpts)) CauchyCorr(distMatrix,rho=0.1,shape=1,longdep=10) # See ?MaternCorr for examples of syntaxes for group-specific random effects, # also handled by Cauchy().
A small data set used by Booth & Hobert (1998).
data("clinics")
data("clinics")
A data frame with 16 observations on the following 4 variables.
npos
a numeric vector
nneg
a numeric vector
treatment
a numeric vector
clinic
a numeric vector
Booth, J.G., Hobert, J.P. (1998) Standard errors of prediction in generalized linear mixed models. J. Am. Stat. Assoc. 93: 262-272.
data(clinics) ## Not run: # The dataset was built as follows npos <- c(11,16,14,2,6,1,1,4,10,22,7,1,0,0,1,6) ntot <- c(36,20,19,16,17,11,5,6,37,32,19,17,12,10,9,7) treatment <- c(rep(1,8),rep(0,8)) clinic <-c(seq(8),seq(8)) clinics <- data.frame(npos=npos,nneg=ntot-npos,treatment=treatment,clinic=clinic) ## End(Not run)
data(clinics) ## Not run: # The dataset was built as follows npos <- c(11,16,14,2,6,1,1,4,10,22,7,1,0,0,1,6) ntot <- c(36,20,19,16,17,11,5,6,37,32,19,17,12,10,9,7) treatment <- c(rep(1,8),rep(0,8)) clinic <-c(seq(8),seq(8)) clinics <- data.frame(npos=npos,nneg=ntot-npos,treatment=treatment,clinic=clinic) ## End(Not run)
The COM-Poisson family is a generalization of the Poisson family which can describe over-dispersed as well as under-dispersed count data. It is indexed by a parameter nu
that quantifies such dispersion. For nu
>1, the distribution is under-dispersed relative to the Poisson distribution with same mean. It includes the Poisson, geometric and Bernoulli as special (or limit) cases (see Details). The COM-Poisson family is here implemented as a family
object, so that it can be fitted by glm
, and further used to model conditional responses in mixed models fitted by this package's functions (see Examples). nu
is distinct from the dispersion parameter considered elsewhere in this package and in the GLM literature, as
affects in a more specific way the log-likelihood.
Several links are now allowed for this family, corresponding to different versions of the COMPoisson described in the literature (e.g., Sellers & Shmueli 2010; Huang 2017).
COMPoisson(nu = stop("COMPoisson's 'nu' must be specified"), link = "loglambda")
COMPoisson(nu = stop("COMPoisson's 'nu' must be specified"), link = "loglambda")
link |
GLM link function. The default is the canonical link |
nu |
Under-dispersion parameter. The |
The th term of the distribution can be written
where
and
, for
implied by its inverse relationship, the expectation formula
. The case
nu=0
is the geometric distribution with parameter ;
nu=1
is the Poisson distribution with mean ; and the limit as
nu
-> is the Bernoulli distribution with expectation
.
From this definition, this is an exponential family model with canonical parameters and
. When the linear predictor
specifies
, the canonical link is used (e.g., Sellers & Shmueli 2010). It is here nicknamed
"loglambda"
and does not have a known expression in terms of elementary functions. To obtain as the link inverse of the linear predictor
, one then first computes
and then
by the expectation formula. For other links (Huang 2017), one directly computes
by the link inverse (e.g.,
for link
"log"
), and then one may solve for to obtain other features of the distribution.
The relationships between and
or other moments of the distribution involve infinite summations. These sums can be easily approximated by a finite number of terms for large
nu
but not when nu
approaches zero. For this reason, the code may fail to fit distributions with nu
approaching 0 (strong residual over-dispersion). The case nu=0
(the geometric distribution) is fitted by an ad hoc algorithm devoid of such problems. Otherwise, spaMM
truncates the sum, and uses numerical integrals to approximate missing terms (which slows down the fitting operation). In addition, it applies an ad hoc continuity correction to ensure continuity of the result in nu=1
(Poisson case). These corrections affect numerical results for the case of residual overdispersion but are negligible for the case of residual underdispersion. Alternatively, spaMM
uses Gaunt et al.'s (2017) approximations when the condition defined by spaMM.getOption("CMP_asympto_cond")
is satisfied. All approximations reduces the accuracy of computations, in a way that can impede the extended Levenberg-Marquardt algorithm sometimes needed by spaMM.
The name COMP_nu
should be used to set initial values or bounds on nu
in control arguments of the fitting functions (e.g., fitme(.,init=list(COMP_nu=1))
). Fixed values should be set by the family argument (COMPoisson(nu=.)
).
A family object.
Gaunt, Robert E. and Iyengar, Satish and Olde Daalhuis, Adri B. and Simsek, Burcin. (2017) An asymptotic expansion for the normalizing constant of the Conway–Maxwell–Poisson distribution. Ann Inst Stat Math doi:10.1007/s10463-017-0629-6.
Huang, Alan (2017) Mean-parametrized Conway-Maxwell-Poisson regression models for dispersed counts. Stat. Modelling doi:10.1177/1471082X17697749
G. Shmueli, T. P. Minka, J. B. Kadane, S. Borle and P. Boatwright (2005) A useful distribution for fitting discrete data: revival of the Conway-Maxwell-Poisson distribution. Appl. Statist. 54: 127-142.
Sellers KF, Shmueli G (2010) A Flexible Regression Model for Count Data. Ann. Appl. Stat. 4: 943–961
if (spaMM.getOption("example_maxtime")>0.9) { # Fitting COMPoisson model with estimated nu parameter: # data("freight") ## example from Sellers & Shmueli, Ann. Appl. Stat. 4: 943–961 (2010) fitme(broken ~ transfers, data=freight, family = COMPoisson()) fitme(broken ~ transfers, data=freight, family = COMPoisson(link="log")) # glm(), HLCor() and HLfit() handle spaMM::COMPoisson() with fixed overdispersion: # glm(broken ~ transfers, data=freight, family = COMPoisson(nu=10)) HLfit(broken ~ transfers+(1|id), data=freight, family = COMPoisson(nu=10),method="ML") # Equivalence of poisson() and COMPoisson(nu=1): # COMPglm <- glm(broken ~ transfers, data=freight, family = poisson()) coef(COMPglm) logLik(COMPglm) COMPglm <- glm(broken ~ transfers, data=freight, family = COMPoisson(nu=1)) coef(COMPglm) logLik(COMPglm) HLfit(broken ~ transfers, data=freight, family = COMPoisson(nu=1)) }
if (spaMM.getOption("example_maxtime")>0.9) { # Fitting COMPoisson model with estimated nu parameter: # data("freight") ## example from Sellers & Shmueli, Ann. Appl. Stat. 4: 943–961 (2010) fitme(broken ~ transfers, data=freight, family = COMPoisson()) fitme(broken ~ transfers, data=freight, family = COMPoisson(link="log")) # glm(), HLCor() and HLfit() handle spaMM::COMPoisson() with fixed overdispersion: # glm(broken ~ transfers, data=freight, family = COMPoisson(nu=10)) HLfit(broken ~ transfers+(1|id), data=freight, family = COMPoisson(nu=10),method="ML") # Equivalence of poisson() and COMPoisson(nu=1): # COMPglm <- glm(broken ~ transfers, data=freight, family = poisson()) coef(COMPglm) logLik(COMPglm) COMPglm <- glm(broken ~ transfers, data=freight, family = COMPoisson(nu=1)) coef(COMPglm) logLik(COMPglm) HLfit(broken ~ transfers, data=freight, family = COMPoisson(nu=1)) }
An example of a composite random effect is corrMatrix(sex|pair)
. It combines features of a random-coefficient model (sex|pair)
and of an autocorrelated random effect corrMatrix(.|.)
. The random-coefficient model is characterized by a covariance matrix C for the random effects
and
both affecting each of the two sexes for each
pair
, and the corrMatrix
random effect assumes that elements of each of the two vectors for
pair
=1,..., are correlated according to a given
correlation matrix A. Then the composite random effect is defined as the one with 2
2
covariance matrix
kronecker
(C,A).
The definition of composite random effects through the kronecker
product may be motivated and understood in light of a quantitative-genetics application (see help("Gryphon") for an example). In this context the two response variables are typically two individual traits. Each trait is affected by two sets of genes, the effect of each set being represented by a gaussian random effect (u_1
or u_2
). The effect of genetic relatedness on the correlation of random effects u_i,ID
among individuals ID
within each set of genes is described by the corrMatrix A. The effects on the two traits for each individual are interpreted as different linear combinations of these two random effects (the coefficients of these linear combinations determining the C matrix). Under these assumptions the correlation matrix of the responses (in order (trait, individual)=(1,1)...(1,ID)... (2,1)...(2,ID)...) is indeed
kronecker
(C,A).
Composite random effects are not restricted to corrMatrix
terms, and can also be fitted for multivariate-response models. For example, Matern(mv(1,2)|longitude+latitude)
terms can be fitted, in which case the correlation model is still defined through the Kronecker product, where A will be a (fitted) Matérn correlation matrix, and C will be the correlation matrix of the fitted random-coefficient model for the mv
virtual factor for multivariate response.
Implementation and tests of composite random effects is work in progress, with the following ones having been tested: corrMatrix
, Matern
, Cauchy
, adjacency
, IMRF
, AR1
, and to a lesser extent MaternIMRFa
. Fits of other composite terms may fail. Even if they succeed, not all post-fit procedures may be operational: in particular, prediction (and then, simulation) with newdata
may not yet work.
Further, as for random-coefficient terms in univariate-response models, some components of the computed prediction variance depend on a poorly characterized approximation, yielding different results for different fitting algorithms (see Details in predVar
).
The summary of the model provides fitted parameters for A if this matrix derives from a parametric correlation model (e.g., Matern
), and a description of the C matrix where it is viewed as a covariance matrix, parametrized by its variances and its correlation coefficient(s). In a standard random-coefficient model these variances would be those of the correlated random effects (see summary.HLfit
). In the composite random-effect model this is not necessarily so as the variance of the correlated random effects also depend on the variances implied by the A matrix, which are not necessarily 1 if A is a covariance matrix rather than simply a correlation matrix.
A <prefix>(<LHS>|<RHS>)
term is *not* a composite random effect when the LHS in boolean, a factor from boolean, or “0+numeric”. See the Matérn examples, and the corrMatrix “<LHS> is 0+numeric
” example, below.
if (spaMM.getOption("example_maxtime")>1.8) { ######################## #### corrMatrix examples ######################## ## Toy data preparation data("blackcap") toy <- blackcap toy$ID <- gl(7,2) grp <- rep(1:2,7) toy$migStatus <- toy$migStatus +(grp==2) toy$loc <- rownames(toy) # to use as levels matching the corrMatrix dimnames toy$grp <- factor(grp) toy$bool <- toy$grp==1L toy$boolfac <- factor(toy$bool) toy$num <- seq(from=1, to=2, length.out=14) ## Build a toy corrMatrix as perturbation of identity matrix: n_rhs <- 14L eps <- 0.1 set.seed(123) rcov <- ((1-eps)*diag(n_rhs)+eps*rWishart(1,n_rhs,diag(n_rhs)/n_rhs)[,,1]) # eigen(rcov)$values colnames(rcov) <- rownames(rcov) <- toy$loc # DON'T FORGET NAMES ##### Illustrating the different LHS types ### <LHS> is logical (TRUE/FALSE) => No induced random-coefficient C matrix; # corrMatrix affects only responses for which <LHS> is TRUE: # (fit1 <- fitme(migStatus ~ bool + corrMatrix(bool|loc), data=toy, corrMatrix=rcov)) # # Matrix::image(get_ZALMatrix(fit1)) ### <RHS> is a factor built from a logical => same a 'logical' case above: # (fit2 <- fitme(migStatus ~ boolfac + corrMatrix(boolfac|loc), data=toy, corrMatrix=rcov)) # # Matrix::image(get_ZALMatrix(fit2)) ### <RHS> is a factor not built from a logical: # (grp|.) and (0+grp|.) lead to equivalent fits of the same composite model, # but contrasts are not used in the second case and the C matrices differ, # as for standard random-coefficient models. # (fit1 <- fitme(migStatus ~ grp + corrMatrix(grp|loc), data=toy, corrMatrix=rcov)) (fit2 <- fitme(migStatus ~ grp + corrMatrix(0+grp|loc), data=toy, corrMatrix=rcov)) # # => same fits, but different internal structures: Matrix::image(fit1$ZAlist[[1]]) # (contrasts used) Matrix::image(fit2$ZAlist[[1]]) # (contrasts not used) # Also compare ranef(fit1) versus ranef(fit2) # # ## One can fix the C matrix, as for standard random-coefficient terms # (fit1 <- fitme(migStatus ~ grp + corrMatrix(0+grp|loc),data=toy, corrMatrix=rcov, fixed=list(ranCoefs=list("1"=c(1,0.5,1))))) # # same result without contrasts hence different 'ranCoefs': # (fit2 <- fitme(migStatus ~ grp + corrMatrix(grp|loc), data=toy, corrMatrix=rcov, fixed=list(ranCoefs=list("1"=c(1,-0.5,1))))) ### <LHS> is numeric (but not '0+numeric'): # composite model with C being 2*2 for Intercept and numeric variable # (fitme(migStatus ~ num + corrMatrix(num|loc), data=toy, corrMatrix=rcov)) ### <LHS> is 0+numeric: no random-coefficient C matrix # as the Intercept is removed, but the correlated random effects # arising from the corrMatrix are multiplied by sqrt(<numeric variable>) # (fitme(migStatus ~ num + corrMatrix(0+num|loc), data=toy, corrMatrix=rcov)) ### <LHS> for multivariate response (see help("Gryphon") for more typical example) ## More toy data preparation for multivariate response ch <- chol(rcov) set.seed(123) v1 <- tcrossprod(ch,t(rnorm(14,sd=1))) v2 <- tcrossprod(ch,t(rnorm(14,sd=1))) toy$status <- 2*v1+v2 toy$status2 <- 2*v1-v2 ## Fit: fitmv(submodels=list(mod1=list(status ~ 1+ corrMatrix(0+mv(1,2)|loc)), mod2=list(status2 ~ 1+ corrMatrix(0+mv(1,2)|loc))), data=toy, corrMatrix=rcov) ################################################## #### Matern examples: sex-dependent random effects ################################################## if (spaMM.getOption("example_maxtime")>2) { data(Leuca) subLeuca <- Leuca[c(1:10,79:88),] # subset for faster examples # The random effects in the following examples are composite because 'sex' is not # boolean nor factor from boolean. If 'Matern(factor(female)|x+y)' were used, the effect # would be the same 'Matern(female|x)', fitting fitme(fec_div ~ 1 + Matern(sex|x+y),data=subLeuca) # => builds a random effect # correlated across sexes, from 2 independent realizations u_1 and u_2 of 20 values # (for the 20 locations in the data). In a (sex|x) term the 20 values would be # independent from each other within each u_i. In the Matern(sex|x+y) such 20 values # are autocorrelated within each u_i. # For pedagogic purposes, one can also fit fitme(fec_div ~ 1 + Matern(sex|x + y %in% sex),data=subLeuca) # which again builds a random effect from 2 independent realizations # u_1 and u_2, but each u_i now contains two realizations u_i1 and u_i2 of 10 values, # autocorrelated within each u_ij following the Matern model, # but independent between u_i1 and u_i2. As a result, the overall random effect in each sex, # v_m or v_f, is a weighted sum of two sex-specific Matern random effect, # so that v_m and v_f are independent from each other. } }
if (spaMM.getOption("example_maxtime")>1.8) { ######################## #### corrMatrix examples ######################## ## Toy data preparation data("blackcap") toy <- blackcap toy$ID <- gl(7,2) grp <- rep(1:2,7) toy$migStatus <- toy$migStatus +(grp==2) toy$loc <- rownames(toy) # to use as levels matching the corrMatrix dimnames toy$grp <- factor(grp) toy$bool <- toy$grp==1L toy$boolfac <- factor(toy$bool) toy$num <- seq(from=1, to=2, length.out=14) ## Build a toy corrMatrix as perturbation of identity matrix: n_rhs <- 14L eps <- 0.1 set.seed(123) rcov <- ((1-eps)*diag(n_rhs)+eps*rWishart(1,n_rhs,diag(n_rhs)/n_rhs)[,,1]) # eigen(rcov)$values colnames(rcov) <- rownames(rcov) <- toy$loc # DON'T FORGET NAMES ##### Illustrating the different LHS types ### <LHS> is logical (TRUE/FALSE) => No induced random-coefficient C matrix; # corrMatrix affects only responses for which <LHS> is TRUE: # (fit1 <- fitme(migStatus ~ bool + corrMatrix(bool|loc), data=toy, corrMatrix=rcov)) # # Matrix::image(get_ZALMatrix(fit1)) ### <RHS> is a factor built from a logical => same a 'logical' case above: # (fit2 <- fitme(migStatus ~ boolfac + corrMatrix(boolfac|loc), data=toy, corrMatrix=rcov)) # # Matrix::image(get_ZALMatrix(fit2)) ### <RHS> is a factor not built from a logical: # (grp|.) and (0+grp|.) lead to equivalent fits of the same composite model, # but contrasts are not used in the second case and the C matrices differ, # as for standard random-coefficient models. # (fit1 <- fitme(migStatus ~ grp + corrMatrix(grp|loc), data=toy, corrMatrix=rcov)) (fit2 <- fitme(migStatus ~ grp + corrMatrix(0+grp|loc), data=toy, corrMatrix=rcov)) # # => same fits, but different internal structures: Matrix::image(fit1$ZAlist[[1]]) # (contrasts used) Matrix::image(fit2$ZAlist[[1]]) # (contrasts not used) # Also compare ranef(fit1) versus ranef(fit2) # # ## One can fix the C matrix, as for standard random-coefficient terms # (fit1 <- fitme(migStatus ~ grp + corrMatrix(0+grp|loc),data=toy, corrMatrix=rcov, fixed=list(ranCoefs=list("1"=c(1,0.5,1))))) # # same result without contrasts hence different 'ranCoefs': # (fit2 <- fitme(migStatus ~ grp + corrMatrix(grp|loc), data=toy, corrMatrix=rcov, fixed=list(ranCoefs=list("1"=c(1,-0.5,1))))) ### <LHS> is numeric (but not '0+numeric'): # composite model with C being 2*2 for Intercept and numeric variable # (fitme(migStatus ~ num + corrMatrix(num|loc), data=toy, corrMatrix=rcov)) ### <LHS> is 0+numeric: no random-coefficient C matrix # as the Intercept is removed, but the correlated random effects # arising from the corrMatrix are multiplied by sqrt(<numeric variable>) # (fitme(migStatus ~ num + corrMatrix(0+num|loc), data=toy, corrMatrix=rcov)) ### <LHS> for multivariate response (see help("Gryphon") for more typical example) ## More toy data preparation for multivariate response ch <- chol(rcov) set.seed(123) v1 <- tcrossprod(ch,t(rnorm(14,sd=1))) v2 <- tcrossprod(ch,t(rnorm(14,sd=1))) toy$status <- 2*v1+v2 toy$status2 <- 2*v1-v2 ## Fit: fitmv(submodels=list(mod1=list(status ~ 1+ corrMatrix(0+mv(1,2)|loc)), mod2=list(status2 ~ 1+ corrMatrix(0+mv(1,2)|loc))), data=toy, corrMatrix=rcov) ################################################## #### Matern examples: sex-dependent random effects ################################################## if (spaMM.getOption("example_maxtime")>2) { data(Leuca) subLeuca <- Leuca[c(1:10,79:88),] # subset for faster examples # The random effects in the following examples are composite because 'sex' is not # boolean nor factor from boolean. If 'Matern(factor(female)|x+y)' were used, the effect # would be the same 'Matern(female|x)', fitting fitme(fec_div ~ 1 + Matern(sex|x+y),data=subLeuca) # => builds a random effect # correlated across sexes, from 2 independent realizations u_1 and u_2 of 20 values # (for the 20 locations in the data). In a (sex|x) term the 20 values would be # independent from each other within each u_i. In the Matern(sex|x+y) such 20 values # are autocorrelated within each u_i. # For pedagogic purposes, one can also fit fitme(fec_div ~ 1 + Matern(sex|x + y %in% sex),data=subLeuca) # which again builds a random effect from 2 independent realizations # u_1 and u_2, but each u_i now contains two realizations u_i1 and u_i2 of 10 values, # autocorrelated within each u_ij following the Matern model, # but independent between u_i1 and u_i2. As a result, the overall random effect in each sex, # v_m or v_f, is a weighted sum of two sex-specific Matern random effect, # so that v_m and v_f are independent from each other. } }
This function interfaces two procedures: a profile confidence interval procedure implemented for fixed-effects coefficients only; and a parametric bootstrap procedure that can be used to provide confidence interval for any parameter, whether a canonical parameter of the model or any function of one or several such parameters.
The bootstrap is performed if the parm
argument is a function or a quoted expression or if the boot_args
argument is a list. The profile confidence interval is computed if neither of these conditions is true. In that case parm
must be the name(s) of some fixed-effect coefficient, and the (p_v
approximation of the) profile likelihood ratio for the given parameter is used to define the interval, where the profiling is over all other fitted parameters, including other fixed-effects coefficients, as well as variances of random effects and spatial correlations if these were fitted.
Of related interest, see numInfo
which evaluates numerically the information matrix for given sets of canonical model parameters, from which asymptotic confidence intervals can be deduced.
## S3 method for class 'HLfit' confint(object, parm, level=0.95, verbose=TRUE, boot_args=NULL, format="default", ...)
## S3 method for class 'HLfit' confint(object, parm, level=0.95, verbose=TRUE, boot_args=NULL, format="default", ...)
object |
An object of class |
parm |
character vector, integer vector, or function, or a quoted expression. If character, the name(s) of parameter(s) to be fitted; if integer, their position in the |
level |
The coverage of the interval. |
verbose |
whether to print the interval or not. As the function returns its more extensive results invisibly, this printing is the only visible output. |
boot_args |
NULL or a list of arguments passed to functions |
format |
Only effective non-default value is |
... |
Additional arguments (maybe not used, but conforming to the generic definition of |
The format of the value varies, but in all cases distinguished below, one or more tables are included, as a table
attribute, in the format of the stats::confint
result, to facilitate consistent extraction of results. By default confint
returns invisibly the full values described below, but if format="stats"
, only the table
attribute is returned.
If a profile CI has been computed for a single parameter, a list is returned including the confidence interval
as shown by verbose=TRUE
, and the fits lowerfit
and upperfit
giving the profile fits at the confidence bounds. This list bears the table
attribute.
If a profile CI has been computed for several parameters, a structured list, named according to the parameter names, of such single-parameter results is returned, and a single table
attribute for all parameters is attached to the upper level of the list.
If a bootstrap was performed, for a single parameter the result of the boot.ci
call is returned, to which a table
attribute is added. This table
is now a list of tables for the different bootstrap CI types (default being normal
, percent
, and basic
), each such table in the format of the stats::confint
results. For several parameters, a named list of boot.ci
results is returned, its names being the parameter names, and the table
attribute is attached to the upper level of the list.
The boot.ci
return value for each parameter includes the call to boot.ci
. This call is typically shown including a long t
vector, which makes a bulky display. spaMM hacks the printing to abbreviate long t
s.
numInfo
for information matrix.
data("wafers") wfit <- HLfit(y ~X1+(1|batch), family=Gamma(log), data=wafers, method="ML") confint(wfit,"X1") # profile CI if (spaMM.getOption("example_maxtime")>30) { # bootstrap CI induced by 'boot_args': confint(wfit,names(fixef(wfit)), boot_args=list(nsim=99, seed=123)) # bootstrap CI induced by 'parm' being a function: confint(wfit,parm=function(v) fixef(v), boot_args=list(nb_cores=10, nsim=199, seed=123)) # Same effect if 'parm' is a quoted expression in terms of 'hlfit': confint(wfit,parm=quote(fixef(hlfit)), boot_args=list(nb_cores=10, nsim=199, seed=123)) # CI for the variance of the random effect: ci <- confint(wfit,parm=function(fit){get_ranPars(fit)$lambda[1]}, boot_args=list(nb_cores=10, nsim=199, seed=123)) # The distribution of bootstrap replicates: plot(ecdf(ci$call$t)) # We may be far from ideal condition for accuracy of bootstrap intervals; # for variances, a log transformation may sometimes help, but not here. # Passing arguments to child processes, as documented in help("spaMM_boot"): set.seed(123) rvar <- runif(nrow(wafers)) wrfit <- fitme(y ~X1+(1|batch), family=Gamma(log), data=wafers, fixed=list(phi=rvar)) confint(wrfit, parm = "(Intercept)", boot_args = list(nsim = 100, nb_cores = 2, fit_env = list(rvar=rvar))) }
data("wafers") wfit <- HLfit(y ~X1+(1|batch), family=Gamma(log), data=wafers, method="ML") confint(wfit,"X1") # profile CI if (spaMM.getOption("example_maxtime")>30) { # bootstrap CI induced by 'boot_args': confint(wfit,names(fixef(wfit)), boot_args=list(nsim=99, seed=123)) # bootstrap CI induced by 'parm' being a function: confint(wfit,parm=function(v) fixef(v), boot_args=list(nb_cores=10, nsim=199, seed=123)) # Same effect if 'parm' is a quoted expression in terms of 'hlfit': confint(wfit,parm=quote(fixef(hlfit)), boot_args=list(nb_cores=10, nsim=199, seed=123)) # CI for the variance of the random effect: ci <- confint(wfit,parm=function(fit){get_ranPars(fit)$lambda[1]}, boot_args=list(nb_cores=10, nsim=199, seed=123)) # The distribution of bootstrap replicates: plot(ecdf(ci$call$t)) # We may be far from ideal condition for accuracy of bootstrap intervals; # for variances, a log transformation may sometimes help, but not here. # Passing arguments to child processes, as documented in help("spaMM_boot"): set.seed(123) rvar <- runif(nrow(wafers)) wrfit <- fitme(y ~X1+(1|batch), family=Gamma(log), data=wafers, fixed=list(phi=rvar)) confint(wrfit, parm = "(Intercept)", boot_args = list(nsim = 100, nb_cores = 2, fit_env = list(rvar=rvar))) }
A list of parameters controlling the HLfit
fitting algorithm (potentially called by all fitting functions in spaMM), which should mostly be ignored in routine use. Possible controls are:
algebra, sparse_precision
: see algebra
;
conv.threshold
and spaMM_tol
: spaMM_tol
is a list of tolerance values, with elements Xtol_rel
and Xtol_abs
that define thresholds for relative and absolute changes in parameter values in iterative algorithms (used in tests of the form “d(param)< Xtol_rel * param + Xtol_abs”, so that Xtol_abs
is operative only for small parameter values). conv.threshold
is the older way to control Xtol_rel
. Default values are given by spaMM.getOption("spaMM_tol");
break_conv_logL
: a boolean specifying whether the iterative algorithm should terminate when log-likelihood appears to have converged (roughly, when its relative variation over on iteration is lower than 1e-8). Default is FALSE (convergence is then assessed on the parameter estimates rather than on log-likelihood).
iter.mean.dispFix
: the number of iterations of the iterative algorithm for coefficients of the linear predictor,
if no dispersion parameters are estimated by the iterative algorithm. Defaults to 200 except for Gamma(log)-family models;
iter.mean.dispVar
: the number of iterations of the iterative algorithm for coefficients of the linear predictor,
if some dispersion parameter(s) is estimated by the iterative algorithm. Defaults to 50 except for Gamma(log)-family models;
max.iter
: the number of iterations of the iterative algorithm for joint estimation of dispersion parameters and
of coefficients of the linear predictor. Defaults to 200. This is typically much more than necessary,
unless there is little information to separately estimate and
parameters;
resid.family
: was a previously documented control (before version 2.6.40), and will still operate as previously documented, but should not be used in new code.
# <fitting function>(., control.HLfit=list(...))
# <fitting function>(., control.HLfit=list(...))
spaMM fits can produce various convergence warnings or messages.
Messages referring to convergence issues in initialization can generally be ignored but may help to diagnose other apparent problems, if any.
Warnings referring to .calc_dispGammaGLM
(for residual-dispersion fits) can generally be ignored when they show a small criterion (say <1e-5) but may otherwise suggest that the final fit did not optimize its objective.
Messages pointing to slow convergence and drawing users to this doscumentation do not necessarily mean the fit is incorrect. Rather, they suggest that another fitting strategy could be tried. Keep in mind that several parameters (notably the dispersion parameters: the variance of random effects and the residual variance parameter, if any) can be estimated either by the iterative algorithms, or by generic optimization methods. In my experience, slow convergence happens in certain cases where a large random-effect variance is considered by the algorithm used.
How to know which algorithm has been selected for each parameter? fitme(., verbose=c(TRACE=TRUE))
shows successive values of the variables estimated by optimization (See Examples; if no value appears, then all are estimated by iterative methods). The first lines of the summary of a fit object should tell which variances are estimated by the “outer” method.
If the iterative algorithm is being used, then it is worth trying to use the generic optimization methods. In particular, if you used HLfit
, try using fitme
; if you already use fitme
, try to enforce generic optimization of the random-effect variance(s) (see inits
). Conversely, if generic optimization is being used, the maximum lambda value could be controlled (say, upper=list(lambda=c(10,NA))
), or the iterative algorithm can be called (see inits
again).
For the largest datasets, it may be worth comparing the speed of the "spcorr"
and "spprec"
choices of the algebra
control, in case spaMM has not selected the most appropriate by default. However, this will not be useful for geostatistical models with many spatial locations.
# See help("inits") for examples of control by initial values.
# See help("inits") for examples of control by initial values.
This implements the Matérn correlation structure (see Matern
) for use with lme
or glmmPQL
. Usage is as for others corSpatial
objects
such as corGaus
or corExp
, except that the Matérn family
has an additional parameter. This function was defined for comparing
results obtained with corrHLfit
to those produced by lme
and glmmmPQL
. There
are problems in fitting (G)LMMs in the latter way, so it is not a
recommended practice.
corMatern(value = c(1, 0.5), form = ~1, nugget = FALSE, nuScaled = FALSE, metric = c("euclidean", "maximum", "manhattan"), fixed = FALSE)
corMatern(value = c(1, 0.5), form = ~1, nugget = FALSE, nuScaled = FALSE, metric = c("euclidean", "maximum", "manhattan"), fixed = FALSE)
value |
An optional vector of parameter values, with serves as initial values or as fixed values depending on the If If |
form |
(Pasted from corSpatial) a one sided formula of the form |
nugget |
an optional logical value indicating whether a nugget
effect is present. Defaults to |
nuScaled |
If |
metric |
(Pasted from corSpatial) an optional character string specifying the distance
metric to be used. The currently available options are
|
fixed |
an optional logical value indicating whether the
coefficients should be allowed to vary in the optimization, or kept
fixed at their initial value. Defaults to |
This function is a constructor for the corMatern
class,
representing a Matérn spatial correlation structure. See
MaternCorr
for details on the Matérn family.
an object of class corMatern
, also inheriting from class
corSpatial
, representing a Matérn spatial correlation
structure.
The R and C code for the methods for corMatern
objects builds on code for corSpatial
objects, by D.M. Bates, J.C. Pinheiro and S. DebRoy, in a circa-2012 version of nlme.
Mixed-Effects Models in S and S-PLUS, José C. Pinheiro and Douglas M. Bates, Statistics and Computing Series, Springer-Verlag, New York, NY, 2000.
## LMM data("blackcap") blackcapD <- cbind(blackcap,dummy=1) ## obscure, isn't it? ## With method= 'ML' in lme, The correlated random effect is described ## as a correlated residual error and no extra residual variance is fitted: nlme::lme(fixed = migStatus ~ means, data = blackcapD, random = ~ 1 | dummy, correlation = corMatern(form = ~ longitude+latitude | dummy), method = "ML", control=nlme::lmeControl(sing.tol=1e-20)) ## Binomial GLMM if (spaMM.getOption("example_maxtime")>32) { data("Loaloa") LoaloaD <- cbind(Loaloa,dummy=1) MASS::glmmPQL(fixed =cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI, data = LoaloaD, random = ~ 1 | dummy,family=binomial, correlation = corMatern(form = ~ longitude+latitude | dummy)) }
## LMM data("blackcap") blackcapD <- cbind(blackcap,dummy=1) ## obscure, isn't it? ## With method= 'ML' in lme, The correlated random effect is described ## as a correlated residual error and no extra residual variance is fitted: nlme::lme(fixed = migStatus ~ means, data = blackcapD, random = ~ 1 | dummy, correlation = corMatern(form = ~ longitude+latitude | dummy), method = "ML", control=nlme::lmeControl(sing.tol=1e-20)) ## Binomial GLMM if (spaMM.getOption("example_maxtime")>32) { data("Loaloa") LoaloaD <- cbind(Loaloa,dummy=1) MASS::glmmPQL(fixed =cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI, data = LoaloaD, random = ~ 1 | dummy,family=binomial, correlation = corMatern(form = ~ longitude+latitude | dummy)) }
corr_family
objects
corr_family
objects provide a convenient way to implement correlation models handled by spaMM
, analogous to family
objects. These objects are undocumented (but there are documentation pages for each of the models implemented).
# Matern(...) # see help(Matern) # Cauchy(...) # see help(Cauchy) # corrMatrix(...) # see help(corrMatrix) # AR1(...) # see help(AR1) # adjacency(...) # see help(adjacency) # IMRF(...) # see help(IMRF) ## S3 method for class 'corr_family' print(x,...)
# Matern(...) # see help(Matern) # Cauchy(...) # see help(Cauchy) # corrMatrix(...) # see help(corrMatrix) # AR1(...) # see help(AR1) # adjacency(...) # see help(adjacency) # IMRF(...) # see help(IMRF) ## S3 method for class 'corr_family' print(x,...)
x |
|
... |
arguments that may be needed by some |
One can declare and fit correlated random effects belonging to a user-defined correlation (or covariance) model (i.e., a parametric family of correlation matrices, although degenerate case with no parameter are also possible). This documentation is a first introduction to this feature. It is experimental in the sense that its design has been formalized only from a limited number of corrFamily examples, and that the documentation is not mature. Implementing prediction for random-effects defined in this way may be tricky. A distinct documentation corrFamily-design
provides more information for the efficient design of new correlation models to be fitted in this way.
A simple example of random-effect model implemented in this way is the autoregressive model of order (AR(p) in the literature; specifically documented elsewhere, see
ARp
). It can be used as a formula term like other autocorrelated random-effects predefined in spaMM, to be fitted by fitme
or fitmv
:
fitme(lh ~ 1 + ARp(1|time, p=3), # <= declaration of random effect < data and other possible arguments >)
User-defined correlation models should be registered for this simple syntax to work (see Details for an alternative syntax):
myARp <- ARp # 'myARP' is thus a user-defined model register_cF("myARp") # Register it so that the next call works fitme(lh ~ 1 + myARp(1|time, p=3), < data and other possible arguments >)
The ARp
object here copied in myARp
is a function (the corrFamily constructor) which returns a list
(the corrFamily descriptor) which contains the necessary information to fit a random effect with an AR(p) correlation. The p
argument in the myARp(1|time, p=3)
term enforces evaluation of myARp(p=3)
, producing the descriptor for the AR(3) model. The structure of this descriptor is
List of 5 $ Cf :function (parvec) ..- < with some attribute > $ tpar : num [1:3] 0.5 0.333 0.25 $ type : chr "precision" $ initialize :function (Zmatrix, ...) ..- < with some attribute > $ fixed : NULL $ calc_moreargs:function (corrfamily, ...) ..- < with some attribute > $ levels_type : chr "time_series" $ calc_corr_from_dist:function (ranFix, char_rd, distmat, ...) ..- < with some attribute > < and possibly other elements >
The meaning of these elements and some additional ones is explained below.
Only Cf
and tpar
are necessary elements of a corrFamily object. If one designs a new descriptor where some other elements are absent, spaMM will try to provide plausible defaults for these elements. Further, if the descriptor does not provide parameter names (as the names of tpar
, or in some more cryptic way), default names "p1"
, "p2"
... will be provided.
## corrFamily descriptor provided as a list of the form # # list(Cf=<.>, tpar=<.>, ...) ## corrFamily constructor: any function that returns # a valid corrFamily descriptor # # function(tpar=<.>, fixed=<.>, ...) # typical but not mandatory arguments ## There is a distinct documentation page for 'register_cF'.
## corrFamily descriptor provided as a list of the form # # list(Cf=<.>, tpar=<.>, ...) ## corrFamily constructor: any function that returns # a valid corrFamily descriptor # # function(tpar=<.>, fixed=<.>, ...) # typical but not mandatory arguments ## There is a distinct documentation page for 'register_cF'.
Elements of the corrFamily descriptor:
Cf |
(required): function returning the correlation matrix (or covariance matrix, or their inverse), given its first argument, a parameter vector. |
tpar |
(required): a feasible argument of |
type |
optional, but required if the return value of |
fixed |
optional: fixed values for some correlation parameters, provided as a named vector with names consistent with those to be used for |
calc_moreargs |
optional: a function returning a list with possible elements |
initialize |
optional: a function evaluating variables that may be needed repeatedly by |
Af |
This function should be defined if the correlation model requires an A matrix (the middle term in the case the design matrix of a random effect term is described by a triple matrix product ZAL as described in |
levels_type |
In the above example its value |
calc_corr_from_dist , make_new_corr_lists
|
Functions possibly needed for prediction (see Details). |
need_Cnn |
optional: a boolean; default is TRUE. Controls prediction computations (see Details). |
public |
An environment where some variables can be saved, typically by the |
fitting-function arguments:
lower, upper, init
and fixed
optimization controls can be used to control optimization of continuous parameters as for other random-effect parameters. They are specified as numeric vectors, themselves being element of the corrPars
list (see Examples in corrFamily-design
). Parameter names (consistent with those to be used for the tpar
argument) may be required to disambiguate incomplete vectors (e.g., to specify only its second element). Apart from fixed
ones, any of the values not specified through the fitting-function arguments will be sought in the return value of the calc_moreargs
function, if provided in the descriptor. If the lower
or upper
information is missing there, it must be provided throught the fitting-function call. If the init
information is missing, a default value will be deduced from the bounds. The init
specification is thus always optional while the bounds specification is optional only if the descriptor provides default values.
Arguments of the corrFamily constructor
These may be ad libitum, as design rules are defined only for the returned descriptor. However, arguments tpar
, fixed
, and public
of predefined constructors, such as ARp
, are designed to match the above respective elements of the descriptor.
* Constructor elements for prediction:
For prediction of autocorrelated random effects, one must first assess whether levels of the random effect not represented in the fit are possible in new data (corresponding to new spatial locations in geostatistical models, or new time steps in time series). In that case need_Cnn
must be TRUE (Interpolated MRFs do not require this as all required random-effect levels are determined by the IMRF mesh
argument rather than by the fitted data or new data).
Next, for autocorrelated random effects where need_Cnn
is TRUE, a make_new_corr_lists
function must be provided, except when a calc_corr_from_dist
function is instead provided (which may be sufficient for models that construct the correlation from a spatial distance matrix). When need_Cnn
is FALSE, a make_new_corr_lists
function may still be needed.
The Examples section provides a simple example of such design, and the source code of the ARp
or ARMA
constructors provide further examples. They show that the make_new_corr_lists
function may assign matrices or vectors as elements of several lists contained in a newLv_env
environment. A matrix is assigned in the cov_newLv_oldv_list
, specifying correlations between “new” levels of the random effect (implied by the new data) and “old” levels (those already included in the design matrix of the random effect for the fit). If need_Cnn
is TRUE, a second matrix may be assigned in the cov_newLv_newLv_list
, specifying correlation between “new” levels, and the diagonal of this matrix is assigned in the diag_cov_newLv_newLv_list
. The overall structure of the code (the conditions where these assignments are performed, and the list indices), should be conserved.
When calling simulate(., newdata=<non-NULL>, type="marginal"
, a fourth matrix may be useful, assigned into a L_newLv_newLv_list
, specifying the matrix root (as a tcrossprod
factor) of the correlation matrix stored in cov_newLv_newLv_list
. The relevant spaMM procedure will however try to compute it on the fly when it has not been provided by the make_new_corr_lists
function.
* Fitting a corrFamily
without a constructor:
It is possible to use an unregistered corrFamily, as follows:
AR3 <- ARp(p=3) # Generate descriptor of AR model of order 3 fitme(lh ~ 1 + corrFamily(1|time), # <= declaration of random effect covStruct=list( corrFamily= AR3 # <= declaration of its correlation structure ), < data and other possible arguments >)
Here the fit only uses a descriptor list, not a constructor function. This descriptor is here provided to the fitting function as an element of the covStruct
argument (using the general syntax of this argument), and in the model formula the corresponding random effect is declared as a term of the formcorrFamily(1|<grouping factor>)
.
This syntax is more complex than the one using a registered constructor, but it might be useful for development purposes (one only has to code the descriptor, not the constructor function). However, it is not general; in particular, using registered constructors may be required to obtain correct results when fitting multivariate-response models by fitmv
.
See ARp
, diallel
, and MaternIMRFa
for basic examples of using a predefined corrFamily descriptor, and corrFamily-design
for more geeky stuff including examples of implementing simple new correlation families.
## Not run: ### Minimal (with many features missing) reimplementation # of corrMatrix() terms as a corrFamily corrMatrix_cF <- function(corrMatrix) { force(corrMatrix) # Makes it available in the environment of the functions next defined. oldZlevels <- NULL initialize <- function(Zmatrix, ...) { oldZlevels <<- colnames(Zmatrix) # Pass info about levels of the random effect in the data. } Cf <- function(newlevels=oldZlevels ) { if (length(newlevels)) { corrMatrix[newlevels,newlevels] } else corrMatrix[oldZlevels,oldZlevels] # for Cf(tpar=numeric(0L)) } calc_moreargs <- function(corrfamily, ...) { list(init=c(),lower=c(),upper=c()) } make_new_corr_lists <- function(newLv_env, which_mats, newZAlist, new_rd, ...) { newlevels <- colnames(newZAlist[[new_rd]]) newLv_env$cov_newLv_oldv_list[[new_rd]] <- corrMatrix[newlevels,oldZlevels, drop=FALSE] if (which_mats$nn[new_rd]) { newLv_env$cov_newLv_newLv_list[[new_rd]] <- corrMatrix[newlevels,newlevels, drop=FALSE] } else { newLv_env$diag_cov_newLv_newLv_list[[new_rd]] <- rep(1,length(newlevels)) } } list(Cf=Cf, tpar=numeric(0L), initialize=initialize, calc_moreargs=calc_moreargs, make_new_corr_lists=make_new_corr_lists, tag="corrMatrix_cF") } register_cF("corrMatrix_cF") # usage: data("blackcap") MLcorMat <- MaternCorr(proxy::dist(blackcap[,c("latitude","longitude")]), nu=0.6285603,rho=0.0544659) corrmat <- proxy::as.matrix(MLcorMat, diag=1) fitme(migStatus ~ means+ corrMatrix_cF(1|name, corrMatrix=corrmat),data=blackcap, corrMatrix=MLcorMat,method="ML") unregister_cF("corrMatrix_cF") # Tidy things before leaving. ## End(Not run)
## Not run: ### Minimal (with many features missing) reimplementation # of corrMatrix() terms as a corrFamily corrMatrix_cF <- function(corrMatrix) { force(corrMatrix) # Makes it available in the environment of the functions next defined. oldZlevels <- NULL initialize <- function(Zmatrix, ...) { oldZlevels <<- colnames(Zmatrix) # Pass info about levels of the random effect in the data. } Cf <- function(newlevels=oldZlevels ) { if (length(newlevels)) { corrMatrix[newlevels,newlevels] } else corrMatrix[oldZlevels,oldZlevels] # for Cf(tpar=numeric(0L)) } calc_moreargs <- function(corrfamily, ...) { list(init=c(),lower=c(),upper=c()) } make_new_corr_lists <- function(newLv_env, which_mats, newZAlist, new_rd, ...) { newlevels <- colnames(newZAlist[[new_rd]]) newLv_env$cov_newLv_oldv_list[[new_rd]] <- corrMatrix[newlevels,oldZlevels, drop=FALSE] if (which_mats$nn[new_rd]) { newLv_env$cov_newLv_newLv_list[[new_rd]] <- corrMatrix[newlevels,newlevels, drop=FALSE] } else { newLv_env$diag_cov_newLv_newLv_list[[new_rd]] <- rep(1,length(newlevels)) } } list(Cf=Cf, tpar=numeric(0L), initialize=initialize, calc_moreargs=calc_moreargs, make_new_corr_lists=make_new_corr_lists, tag="corrMatrix_cF") } register_cF("corrMatrix_cF") # usage: data("blackcap") MLcorMat <- MaternCorr(proxy::dist(blackcap[,c("latitude","longitude")]), nu=0.6285603,rho=0.0544659) corrmat <- proxy::as.matrix(MLcorMat, diag=1) fitme(migStatus ~ means+ corrMatrix_cF(1|name, corrMatrix=corrmat),data=blackcap, corrMatrix=MLcorMat,method="ML") unregister_cF("corrMatrix_cF") # Tidy things before leaving. ## End(Not run)
Tentative formal rules for definition of corrFamily descriptors (work in progress). This is likely to repeat and extend information partially given in corrFamily
and corrFamily-design
documentations.
User-level rules (not relevant fo corrFamily descriptors internally modified during a fit):
Should always be present. For trivial parameterless cases (e.g. ranGCA
), it should be numeric(0L)
, not NULL
.
function; should always be present. For trivial uncorrelated random effects (e.g. ranGCA
,
where only the Af
function carries the information for the model),
it should return an identity matrix, not NULL
,
with row names to be matched to the column names of the Z matrix for the random effect.
optional function. If present, it should have formal arguments including at least corrfamily
and ....
function; optional. If present, it should have row names to be matched to the column names of the Z matrix for the random effect, and also needs column names if it is to be matched with the row names of a correlation matrix (or its inverse).
Optional function. If present, should have formal arguments including at least Zmatrix
and ....
In predefined corrFamily constructors, variables created by initialize
for use by Cf
or Af
should be declared (typically as NULL
) in the body of the constructor, so that R CMD check does not complain.
An environment. initialize
may write into it. It might also read into it, for example read the result of a long previous computation by initialize
during a previous fit, though this opens the door to various errors.
This documentation describe additional design features to be taken into account when defining a new corrFamily
descriptor for a correlation model. Using such a descriptor will be more efficient than the equally general method, of maximizing an objective function of the correlation parameters that calls (say) fitme()
on a model including a corrMatrix
itself function of the correlation parameters. But this may still be inefficient if a few issues are ignored.
For elements of the corrFamily descriptor for basic cases:
The function value should
(1) be of constant class for all parameter values. For families of mathematically sparse matrices, the CsparseMatrix
class is recommended (and more specifically the dsCMatrix
class since the matrix is symmetric); (2) have row names that match the levels of the grouping factor (the nested random effect Example shows the code needed when this nested effect is defined from two variables).
In order to favor the automatic selection of suitable algorithms, tpar
should be chosen so that Cf(tpar)
is least sparse (i.e., has the minimal number of elements equal to zero) in the correlation family, in terms of its sparsity and of the sparsity of its inverse. A tpar
yielding an identity matrix is often a *bad* template as least sparse correlation matrices and their inverses are denser for most families except diagonal ones. For degerate corrFamily objects that describe a constant correlation model rather than a parametric family, use tpar=numeric(0)
.
Do not forget type="precision"
it if the return value of Cf
is an inverse correlation matrix rather than a correlation matrix, in which case one should specify .
should have formal arguments including at least corrfamily
and ...
. The source code of ARp
, ARMA
or diallel
shows the expected structure of its return value.
For advanced features of the corrFamily descriptor:
Af
has (minimally) three formal arguments (newdata, term, ...)
. spaMM will call Af
with distinct values of the newdata
argument for the fit, and for predictions for new data. For the curious: the term
argument that will be provided by spaMM to Af
is the formula term for the random effect – an object of class call
, as obtained e.g. by( ~ 1+ corrFamily(1 | longitude + latitude))[[2]][[3]]
–, which will provide the names of the variables that need to be taken from the newdata
to construct the matrix returned by Af
.
spaMM will regularize invalid or nearly-singular correlation or covariance matrices internally if the correlation function has not done so already, but it it better to control this in the correlation function. The regularize
convenience function is available for that purpose, but parametrizations that avoid the need for regularization are even better, since fitting models with nearly-singular correlation matrices is prone to various difficulties (The Toeplitz example below is good to illustrate potential problems but is otherwise poor as it produces non-positive definite matrices; the ARp
constructor illustrates a parametrization that avoids that problem).
Users should make sure that any regularized matrix still belongs to the intended parametric family of matrices, and they should keep in mind that the spaMM output will show the input parameters of the unregularized matrix, not the parameters of the regularized one (e.g., in the Toeplitz example below, the fitted matrix is a regularized Toepliz matrix with slightly different coefficients than the input parameters).
And for efficiency,
Let us repeat that the correlation function should return matrices of constant class, and in sparse format when the matrices are indeed mathematically sparse. For mathematically dense matrices (as in the Toeplitz example below), the dsyMatrix
class may be suitable.
Let us repeat that in order to favor the automatic selection of suitable algorithms, tpar
should be chosen so that Cf(tpar)
is least sparse in the correlation family. For matrices of CsparseMatrix
, a check is implemented to catch wrong choices of tpar
.
For challenging problems (large data, many parameters...) it may pay to optimize a bit the correlation function. The Example of nested effects with heterogenous variance below illustrates a possible trick. In the same cases, It may also pay to try the alternative algebra
ic methods, by first comparing speed of the different methods (control.HLfit=list(algebra= <"spprec"|"spcorr"|"decorr">)
) for given correlation parameter values, rather than to assume that spaMM will find the best method (even if it often does so).
The corrFamily descriptor may optionally contain booleans possiblyDenseCorr
and sparsePrec
to help spaMM select the most appropriate matrix algebraic methods. sparsePrec
should be set to TRUE if sparse-precision methods are expected to be efficient for fitting the random effect. possiblyDenseCorr
should be set to FALSE if the correlation matrix is expected to be sparse, which means here that less than 15% of its elements are non-zero.
if (spaMM.getOption("example_maxtime")>2 && requireNamespace("agridat", quietly = TRUE)) { data("onofri.winterwheat", package="agridat") ##### Fitting a Toeplitz correlation model for temporal correlations ##### # A Toeplitz correlation matrix of dimension d*d has d-1 parameters # (by symmetry, and with 1s on the main diagonal). These d-1 parameters # can be fitted as follows: Toepfn <- function(v) { toepmat <- Matrix::forceSymmetric(toeplitz(c(1,v))) # dsyMatrix # Many of the matrices in this family are not valid correlation matrices; # the regularize() function is handy here: toepmat <- regularize(toepmat, maxcondnum=1e12) # And don't forget the rownames! rownames(toepmat) <- unique(onofri.winterwheat$year) toepmat } (Toepfit <- spaMM::fitme( yield ~ gen + corrFamily(1|year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily=list(Cf=Toepfn, tpar=rep(1e-4,6))), # (Note the gentle warning if one instead uses tpar=rep(0,6) here) lower=list(corrPars=list("1"=rep(-0.999,6))), upper=list(corrPars=list("1"=rep(0.999,6))))) # The fitted matrix is (nearly) singular, and was regularized: eigen(Corr(Toepfit)[[1]])$values # which means that the returned likelihood may be inaccurate, # and also that the actual matrix elements differ from input parameters: Corr(Toepfit)[[1]][1,-1] ### The usual rules for specifying covStruct, 'lower', 'upper' and 'init' apply # here when the corrFamily term is the second random-effect: (Toep2 <- spaMM::fitme( yield ~ 1 + (1|gen) + corrFamily(1|year), data=onofri.winterwheat, method="REML", covStruct=list("1"=NULL, corrFamily=list(Cf=Toepfn, tpar=rep(1e-4,6))), , init=list(corrPars=list("2"=rep(0.1,6))), lower=list(corrPars=list("2"=rep(-0.999,6))), upper=list(corrPars=list("2"=rep(0.999,6))))) ##### Fitting one variance among years per each of 8 genotypes. ##### # First, note that this can be *more efficiently* fitted by another syntax: ### Fit as a constrained random-coefficient model: # Diagonal matrix of NA's, represented as vector for its lower triangle: ranCoefs_for_diag <- function(nlevels) { vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } (by_rC <- spaMM::fitme(yield ~ 1 + (0+gen|year), data=onofri.winterwheat, method="REML", fixed=list(ranCoefs=list("1"=ranCoefs_for_diag(8))))) ### Fit as a corrFamily model: gy_levels <- paste0(gl(8,1,length =56,labels=levels(onofri.winterwheat$gen)),":", gl(7,8,labels=unique(onofri.winterwheat$year))) # A log scale is often suitable for variances, hence is used below; # a correct but crude implementation of the model is diagf <- function(logvar) { corr_map <- kronecker(Matrix::.symDiagonal(n=7),diag(x=exp(logvar))) rownames(corr_map) <- gy_levels corr_map } # but we can minimize matrix operations as follows: corr_map <- Matrix::.symDiagonal(n=8,x=seq(8)) rownames(corr_map) <- unique(onofri.winterwheat$gen) diagf <- function(logvar) { corr_map@x <- exp(logvar)[corr_map@x] corr_map } # (and this returns a dsCMatrix) (by_cF <- spaMM::fitme( yield ~ 1 + corrFamily(1|gen %in% year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily = list(Cf=diagf, tpar=rep(1,8))), fixed=list(lambda=1), # Don't forget to fix this redundant parameter! # init=list(corrPars=list("1"=rep(log(O.1),8))), # 'init' optional lower=list(corrPars=list("1"=rep(log(1e-6),8))), # 'lower' and 'upper' required upper=list(corrPars=list("1"=rep(log(1e6),8))))) # => The 'gen' effect is nested in the 'year' effect and this must be specified in the # right-hand side of corrFamily(1|gen %in% year) so that the design matrix 'Z' for the # random effects to have the correct structure. And then, as for other correlation # structures (say Matern) it should be necessary to specify only the correlation matrix # for a given year, as done above. Should this fail, it is also possible to specify the # correlation matrix over years, as done below. spaMM will automatically detect, from # its size matching the number of columns of Z, that it must be the matrix over years. corr_map <- Matrix::forceSymmetric(kronecker(Matrix::.symDiagonal(n=7),diag(x=seq(8)))) rownames(corr_map) <- gy_levels diagf <- function(logvar) { corr_map@x <- exp(logvar)[corr_map@x] corr_map } # (and this returns a dsCMatrix) (by_cF <- spaMM::fitme( yield ~ 1 + corrFamily(1|gen %in% year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily = list(Cf=diagf, tpar=rep(1,8))), fixed=list(lambda=1), # Don't forget to fix this redundant parameter! # init=list(corrPars=list("1"=rep(log(O.1),8))), # 'init' optional lower=list(corrPars=list("1"=rep(log(1e-6),8))), # 'lower' and 'upper' required upper=list(corrPars=list("1"=rep(log(1e6),8))))) exp(get_ranPars(by_cF)$corrPars[[1]]) # fitted variances }
if (spaMM.getOption("example_maxtime")>2 && requireNamespace("agridat", quietly = TRUE)) { data("onofri.winterwheat", package="agridat") ##### Fitting a Toeplitz correlation model for temporal correlations ##### # A Toeplitz correlation matrix of dimension d*d has d-1 parameters # (by symmetry, and with 1s on the main diagonal). These d-1 parameters # can be fitted as follows: Toepfn <- function(v) { toepmat <- Matrix::forceSymmetric(toeplitz(c(1,v))) # dsyMatrix # Many of the matrices in this family are not valid correlation matrices; # the regularize() function is handy here: toepmat <- regularize(toepmat, maxcondnum=1e12) # And don't forget the rownames! rownames(toepmat) <- unique(onofri.winterwheat$year) toepmat } (Toepfit <- spaMM::fitme( yield ~ gen + corrFamily(1|year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily=list(Cf=Toepfn, tpar=rep(1e-4,6))), # (Note the gentle warning if one instead uses tpar=rep(0,6) here) lower=list(corrPars=list("1"=rep(-0.999,6))), upper=list(corrPars=list("1"=rep(0.999,6))))) # The fitted matrix is (nearly) singular, and was regularized: eigen(Corr(Toepfit)[[1]])$values # which means that the returned likelihood may be inaccurate, # and also that the actual matrix elements differ from input parameters: Corr(Toepfit)[[1]][1,-1] ### The usual rules for specifying covStruct, 'lower', 'upper' and 'init' apply # here when the corrFamily term is the second random-effect: (Toep2 <- spaMM::fitme( yield ~ 1 + (1|gen) + corrFamily(1|year), data=onofri.winterwheat, method="REML", covStruct=list("1"=NULL, corrFamily=list(Cf=Toepfn, tpar=rep(1e-4,6))), , init=list(corrPars=list("2"=rep(0.1,6))), lower=list(corrPars=list("2"=rep(-0.999,6))), upper=list(corrPars=list("2"=rep(0.999,6))))) ##### Fitting one variance among years per each of 8 genotypes. ##### # First, note that this can be *more efficiently* fitted by another syntax: ### Fit as a constrained random-coefficient model: # Diagonal matrix of NA's, represented as vector for its lower triangle: ranCoefs_for_diag <- function(nlevels) { vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } (by_rC <- spaMM::fitme(yield ~ 1 + (0+gen|year), data=onofri.winterwheat, method="REML", fixed=list(ranCoefs=list("1"=ranCoefs_for_diag(8))))) ### Fit as a corrFamily model: gy_levels <- paste0(gl(8,1,length =56,labels=levels(onofri.winterwheat$gen)),":", gl(7,8,labels=unique(onofri.winterwheat$year))) # A log scale is often suitable for variances, hence is used below; # a correct but crude implementation of the model is diagf <- function(logvar) { corr_map <- kronecker(Matrix::.symDiagonal(n=7),diag(x=exp(logvar))) rownames(corr_map) <- gy_levels corr_map } # but we can minimize matrix operations as follows: corr_map <- Matrix::.symDiagonal(n=8,x=seq(8)) rownames(corr_map) <- unique(onofri.winterwheat$gen) diagf <- function(logvar) { corr_map@x <- exp(logvar)[corr_map@x] corr_map } # (and this returns a dsCMatrix) (by_cF <- spaMM::fitme( yield ~ 1 + corrFamily(1|gen %in% year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily = list(Cf=diagf, tpar=rep(1,8))), fixed=list(lambda=1), # Don't forget to fix this redundant parameter! # init=list(corrPars=list("1"=rep(log(O.1),8))), # 'init' optional lower=list(corrPars=list("1"=rep(log(1e-6),8))), # 'lower' and 'upper' required upper=list(corrPars=list("1"=rep(log(1e6),8))))) # => The 'gen' effect is nested in the 'year' effect and this must be specified in the # right-hand side of corrFamily(1|gen %in% year) so that the design matrix 'Z' for the # random effects to have the correct structure. And then, as for other correlation # structures (say Matern) it should be necessary to specify only the correlation matrix # for a given year, as done above. Should this fail, it is also possible to specify the # correlation matrix over years, as done below. spaMM will automatically detect, from # its size matching the number of columns of Z, that it must be the matrix over years. corr_map <- Matrix::forceSymmetric(kronecker(Matrix::.symDiagonal(n=7),diag(x=seq(8)))) rownames(corr_map) <- gy_levels diagf <- function(logvar) { corr_map@x <- exp(logvar)[corr_map@x] corr_map } # (and this returns a dsCMatrix) (by_cF <- spaMM::fitme( yield ~ 1 + corrFamily(1|gen %in% year), data=onofri.winterwheat, method="REML", covStruct=list(corrFamily = list(Cf=diagf, tpar=rep(1,8))), fixed=list(lambda=1), # Don't forget to fix this redundant parameter! # init=list(corrPars=list("1"=rep(log(O.1),8))), # 'init' optional lower=list(corrPars=list("1"=rep(log(1e-6),8))), # 'lower' and 'upper' required upper=list(corrPars=list("1"=rep(log(1e6),8))))) exp(get_ranPars(by_cF)$corrPars[[1]]) # fitted variances }
This was the first function for fitting all spatial models in spaMM, and is still fully functional, but it is recommended to use fitme
which has different defaults and generally selects more efficient fitting methods, and will handle all classes of models that spaMM can fit, including non-spatial ones. corrHLfit
performs the joint estimation of correlation parameters, fixed effect and dispersion parameters.
corrHLfit(formula, data, init.corrHLfit = list(), init.HLfit = list(), ranFix, fixed=list(), lower = list(), upper = list(), objective = NULL, resid.model = ~1, control.dist = list(), control.corrHLfit = list(), processed = NULL, family = gaussian(), method="REML", nb_cores = NULL, weights.form = NULL, ...)
corrHLfit(formula, data, init.corrHLfit = list(), init.HLfit = list(), ranFix, fixed=list(), lower = list(), upper = list(), objective = NULL, resid.model = ~1, control.dist = list(), control.corrHLfit = list(), processed = NULL, family = gaussian(), method="REML", nb_cores = NULL, weights.form = NULL, ...)
formula |
Either a linear model |
data |
A data frame containing the variables in the response and the model formula. |
init.corrHLfit |
An optional list of initial values for correlation and/or dispersion parameters, e.g.
|
init.HLfit |
See identically named |
fixed , ranFix
|
A list similar to |
lower |
An optional (sub)list of values of the parameters specified through |
upper |
Same as |
objective |
For development purpose, not documented (this had a distinct use in the first version of spaMM, but has been deprecated as such). |
resid.model |
See identically named |
control.dist |
See |
control.corrHLfit |
This may be used control the optimizer. See |
for default values.
processed |
For programming purposes, not documented. |
family |
|
method |
Character: the fitting method to be used, such as |
weights.form |
Specification of prior weights by a one-sided formula: use |
nb_cores |
Not yet operative, only for development purposes. Number of cores to use for parallel computations. |
... |
Optional arguments passed to |
For approximations of likelihood, see method
. For the possible structures of random effects, see random-effects
,
By default corrHLfit
will estimate correlation parameters by maximizing the objective
value returned by HLCor
calls wherein the dispersion parameters are estimated jointly with fixed effects for given correlation parameters. If dispersion parameters are specified in init.corrHLfit
, they will also be estimated by maximizing the objective
value, and HLCor calls
will not estimate them jointly with fixed effects. This means that in general the fixed effect estimates may vary depending on init.corrHLfit
when any form of REML correction is applied.
Correctly using corrHLfit
for likelihood ratio tests of fixed effects may then be tricky. It is safe to perform full ML fits of all parameters (using method="ML"
) for such tests (see Examples). The higher level function fixedLRT
is a safe interface for likelihood ratio tests using some form of REML estimation in corrHLfit
.
attr(<fitted object>,"optimInfo")$lower
and ...$upper
gives the lower and upper bounds for optimization of correlation parameters. These are the default values if the user did not provide explicit values. For the adjacency model, the default values are the inverse of the maximum and minimum eigenvalues of the adjMatrix
. For the Matérn model, the default values are not so easily summarized: they are intended to cover the range of values for which there is statistical information to distinguish among them.
The return value of an HLCor
call, with additional attributes. The HLCor
call is evaluated at the estimated correlation parameter values. These values are included in the return object as its $corrPars
member. The attributes added by corrHLfit
include the original call of the function (which can be retrived by getCall
(<fitted object>), and information about the optimization call within corrHLfit
.
See more examples on data set Loaloa
, to compare fit times by corrHLfit
and fitme
.
See fixedLRT
for likelihood ratio tests.
# Example with an adjacency matrix (autoregressive model): if (spaMM.getOption("example_maxtime")>0.7) { corrHLfit(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix,family=poisson(),data=scotlip,method="ML") } #### Examples with Matern correlations ## A likelihood ratio test based on the ML fits of a full and of a null model. if (spaMM.getOption("example_maxtime")>1.4) { data("blackcap") (fullfit <- corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML") ) (nullfit <- corrHLfit(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, method="ML",init.corrHLfit=list(phi=1e-6))) ## p-value: 1-pchisq(2*(logLik(fullfit)-logLik(nullfit)),df=1) }
# Example with an adjacency matrix (autoregressive model): if (spaMM.getOption("example_maxtime")>0.7) { corrHLfit(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix,family=poisson(),data=scotlip,method="ML") } #### Examples with Matern correlations ## A likelihood ratio test based on the ML fits of a full and of a null model. if (spaMM.getOption("example_maxtime")>1.4) { data("blackcap") (fullfit <- corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML") ) (nullfit <- corrHLfit(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, method="ML",init.corrHLfit=list(phi=1e-6))) ## p-value: 1-pchisq(2*(logLik(fullfit)-logLik(nullfit)),df=1) }
corrMatrix
is an argument of HLCor
, of class dist
or matrix
, with can be used if the model formula contains a term of the form corrMatrix(1|<...>)
. It describes a correlation matrix, possibly as a dist
object. A covariance matrix can actually be passed through this argument, but then it must be a full matrix, not a dist
object. The way the rows and columns of the matrix are matched to the rows of the data
depends on the nature of the grouping term <...>
.
The covStruct
argument can be used for the same purpose and is much more general, in particular allowing to specify several correlation matrices.
The simplest case is illustrated in the first two examples below: the grouping term is identical to a single variable which is present in the data
, whose levels match the rownames of the corrMatrix
. As illustrated by the second example, the order of the data does not matter in that case, because the factor levels are used to match the data
rows to the appropriate row and columns of the corrMatrix
. The corrMatrix
may even contain rows (and columns) in excess of the levels of the grouping term, in which case these rows are ignored.
These convenient properties no longer hold when the grouping term is not a single variable from the data
(third example below), or when its levels do not correspond to row names of the matrix. In these cases, (1) no attempt is made to match the data
rows to the row and column names of the corrMatrix
. Such attempt could succeed only if the user had given names to the matrix matching those that the called function could create from the information in the data
, in which case the user should find easier to specify a single variable that can be matched; (2) the order of data
and corrMatrix
matter; Internally, a single factor variable is constructed from all levels of the variables in the grouping term (i.e., from all levels of latitude
and longitude
, in the third example), with levels 1,2,3... that are matched to rows 1,2,3... of the corrMatrix
. Thus the first row of the data is always associated to the first row of the matrix; (3) further, the dimension of the matrix must match the number of levels implied by the grouping term. For example, one might consider the case of 14 response values but of correlations between only 7 levels of a random effect, with two responses for each level. Then the matrix must be of dimension 7x7.
data("blackcap") ## Here we manually reconstruct the correlation matrix ## of the ML fit produced by corrHLfit: MLcorMat <- MaternCorr(proxy::dist(blackcap[,c("longitude","latitude")]), nu=0.6285603,rho=0.0544659) blackcap$name <- as.factor(rownames(blackcap)) # ## (1) Single variable present in the data # HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=MLcorMat,method="ML") ## (2) Same, permuted: still gives correct result # perm <- sample(14) # Permuted matrix (with permuted names) as 'dist' object pmat <- as.matrix(MLcorMat)[perm,perm] HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=as.dist(pmat),method="ML") # # Permuted matrix (with permuted names) as correlation matrix pcorr <- proxy::as.matrix(MLcorMat, diag=1)[perm,perm] HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=pcorr,method="ML") # ## (3) Other grouping terms (note the messages): # HLCor(migStatus ~ means+ corrMatrix(1|longitude+latitude),data=blackcap, corrMatrix=MLcorMat,method="ML")
data("blackcap") ## Here we manually reconstruct the correlation matrix ## of the ML fit produced by corrHLfit: MLcorMat <- MaternCorr(proxy::dist(blackcap[,c("longitude","latitude")]), nu=0.6285603,rho=0.0544659) blackcap$name <- as.factor(rownames(blackcap)) # ## (1) Single variable present in the data # HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=MLcorMat,method="ML") ## (2) Same, permuted: still gives correct result # perm <- sample(14) # Permuted matrix (with permuted names) as 'dist' object pmat <- as.matrix(MLcorMat)[perm,perm] HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=as.dist(pmat),method="ML") # # Permuted matrix (with permuted names) as correlation matrix pcorr <- proxy::as.matrix(MLcorMat, diag=1)[perm,perm] HLCor(migStatus ~ means+ corrMatrix(1|name),data=blackcap, corrMatrix=pcorr,method="ML") # ## (3) Other grouping terms (note the messages): # HLCor(migStatus ~ means+ corrMatrix(1|longitude+latitude),data=blackcap, corrMatrix=MLcorMat,method="ML")
covStruct
is a formal argument of HLCor
, also handled by fitme
and corrHLfit
, that allows one to specify the correlation structure for different types of random effects, It is an alternative to other ad hoc formal arguments such as corrMatrix
or adjMatrix
. It replaces the deprecated function Predictor(...)
which has served as an interface for specifying the design matrices for random effects in early versions of spaMM
.
The main use of covStruct
is to specify the correlation matrix of levels of a given random effect term, or its inverse (a precision matrix). Assuming that the design matrix of each random effect term follows the structure ZAL described in random-effects
, it is thus an indirect way of specifying the “square root” L of the correlation matrix. The optional A factor can also be given by the optional "AMatrices"
attribute of covStruct
.
covStruct
is a list
of matrices with names specifying the type of matrix considered:covStruct=list(corrMatrix=<some matrix>)
or covStruct=list(adjMatrix=<some matrix>)
, where the “corrMatrix” or “adjMatrix” labels are used to specify the type of information provided (accordingly, the names can be repeated: covStruct=list(corrMatrix=<.>,corrMatrix=<.>)
).
NULL
list members may be necessary, e.g.covStruct=list(corrMatrix=<.>,"2"=NULL,corrMatrix=<.>)
)
when correlations matrices are required only for the first and third random effect.
The covariance structure of a corrMatrix(1|<grouping factor>)
formula term can be specified in two ways (see Examples): either by a correlation matrix factor (covStruct=list(corrMatrix=<some matrix>)
), or by a precision matrix factor Q such that the covariance factor is Q
, using the type name
"precision"
: covStruct=list(precision=<some matrix>)
.
The function as_precision
can be used to perform the conversion from correlation information to precision factor (using a crude solve() that may not always be efficient), but fitting functions may also perform such conversions automatically.
"AMatrices"
is a list of matrices. The names of elements of the list does not matter, but the ith A
matrix, and its row names, should match the ith Z matrix, and its column names. This implies that NULL
list members may be necessary, as for the covStruct
list.
as_precision(corrMatrix, condnum=1e12)
as_precision(corrMatrix, condnum=1e12)
corrMatrix |
Correlation matrix, specified as |
condnum |
Numeric: when a standard Cholesky factorization fails, the matrix is regularized so that the regularized matrix has this condition number (in version 3.10.0 this correction has been implemented more exactly than in previous versions). |
covStruct
can also be specified as a list with an optional "types"
attribute, e.g.structure(list(<some matrix>,types="corrMatrix"))
.
as_precision
returns a list with additional class precision
and with single element a symmetric matrix of class dsCMatrix
.
Gryphon
and pedigree
for a type of applications where declaring a precision matrix is useful.
## Not run: data("blackcap") # a 'dist' object can be used to specify a corrMatrix: MLdistMat <- MaternCorr(proxy::dist(blackcap[,c("latitude","longitude")]), nu=0.6285603,rho=0.0544659) # a 'dist' object! blackcap$name <- as.factor(rownames(blackcap)) fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, corrMatrix=MLdistMat) #### Same result by different input and algorithm: fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, covStruct=list(precision=as_precision(MLdistMat))) # Manual version of the same: as_mat <- proxy::as.matrix(MLdistMat, diag=1) prec_mat <- solve(as_mat) ## precision factor matrix fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, covStruct=list(precision=prec_mat)) # Since no correlation parameter is estimated, # HLcor(., method="ML") is here equivalent to fitme() ## End(Not run)
## Not run: data("blackcap") # a 'dist' object can be used to specify a corrMatrix: MLdistMat <- MaternCorr(proxy::dist(blackcap[,c("latitude","longitude")]), nu=0.6285603,rho=0.0544659) # a 'dist' object! blackcap$name <- as.factor(rownames(blackcap)) fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, corrMatrix=MLdistMat) #### Same result by different input and algorithm: fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, covStruct=list(precision=as_precision(MLdistMat))) # Manual version of the same: as_mat <- proxy::as.matrix(MLdistMat, diag=1) prec_mat <- solve(as_mat) ## precision factor matrix fitme(migStatus ~ means + corrMatrix(1|name), data=blackcap, covStruct=list(precision=prec_mat)) # Since no correlation parameter is estimated, # HLcor(., method="ML") is here equivalent to fitme() ## End(Not run)
ranGCA
and diallel
are random-effect structures designed to represent the effet of symmetric interactions between pairs of individuals (order of individuals in the pair does not matter), while antisym
represents anti-symmetric interactions (the effect of reciprocal ordered pairs on the outcome are opposite, as in the so-called Bradley-Terry models). These random-effect structures all account for multiple membership, i.e., the fact that the same individual may act as the first or the second individual among different pairs, or even within one pair if this makes sense).
More formally, the outcome of an interaction between a pair of agents is subject to a symmetric overall random effect
when the effect “on” individual
(or viewed from the perspective of individual
) equals the effect on
:
. This may result from the additive effect of individual random effects
and
:
, but also from non-additive effects
if the interaction term
is itself symmetric (
).
ranGCA
and diallel
effects represent such symmetric effects, additive or non-additive respectively, in a model formula (see Details for the semantic origin of these names and how they can be changed). Conversely, antisymmetry is characterized by and is represented by the
antisym
formula term.
If individual-level random effects of the form (1|ID1)+ (1|ID2) were included in the model formula instead of ranGCA(1|ID1+ID2)
for symmetric additive interactions, this would result in different variances being fitted for each random effect (breaking the assumption of symmetry), and the value of the random effect would differ for an individual whether it appears as a level of the first random effect or of the second (which is also inconsistent with the idea that the random effect represents a property of the individual).
When ranGCA
or antisym
random effects are fitted, the individual effects are inferred. By contrast, when a diallel
random effect is fitted, an autocorrelated random effect is inferred for each unordered pair (no individual effect is inferred), with correlation
between levels for pairs sharing one individual. This correlation parameter is fitted and is constrained by
(see Details).
ranGCA
is equivalent to the case .
diallel
fits can be slow for large data if the correlation matrix is large, as this matrix can have a fair proportion of nonzero elements.
There may also be identifiability issues for variance parameters: in a LMM as shown in the examples, there will be three parameters for the random variation (phi
, lambda
and rho
) but only two can be estimated if only one observation is made for each dyad.
## formula terms: # ranGCA(1| <.> + <.>) # antisym(1| <.> + <.>) # diallel(1| <.> + <.>, tpar, fixed = NULL, public = NULL) ## where the <.> are two factor identifiers, ** whose levels ## must be identical when representing the same individual ** ## corrFamily constructors: ranGCA() # no argument antisym() # no argument diallel(tpar, fixed = NULL, public = NULL)
## formula terms: # ranGCA(1| <.> + <.>) # antisym(1| <.> + <.>) # diallel(1| <.> + <.>, tpar, fixed = NULL, public = NULL) ## where the <.> are two factor identifiers, ** whose levels ## must be identical when representing the same individual ** ## corrFamily constructors: ranGCA() # no argument antisym() # no argument diallel(tpar, fixed = NULL, public = NULL)
tpar |
Numeric: template value of the correlation coefficient for pairs sharing one individual. |
fixed |
NULL or fixed value of the correlation coefficient. |
public |
NULL, or an environment. When an empty environment is provided, a template |
Although the symmetric random-effect structures may be used in many different contexts (including social network analysis, or “round robin” experiments in psychology; another possibly relevant literature keyword here is “multi membership”), their present names refer to the semantics established for diallel experiments (e.g., Lynch & Walsh, 1998, p. 611), because it is not easy to find a more general yet intuitive semantics. If the names ranGCA
and diallel
sound inappropriate for your context of application, you can declare and use an alternative name for them, taking advantage of the fact that they are random-effect structures defined through corrFamily
constructors, which are functions named as the formula term. For example, symAdd(1|ID1+ID2)
can be used in a model formula after the following two steps:
# Define the 'symAdd' corrFamily constructor (a function) by copy: symAdd <- ranGCA # Associate the 'symAdd' function to 'symAdd' formula terms: register_cF("symAdd")
In diallel experiments, one analyzes the phenotypes of offspring from multiple crosses among which the mother in a cross can be the father in another, so this is an example of multiple-membership. The additive genetic effects of each parent's genotypes are described as “general combining abilities” (GCA). In case of non-additive effect, the half-sib covariance is not half the full-sib covariance and this is represented by the interaction described as “specific combining abilities” (SCA). The sum of GCA and SCA defines a synthetic random effect “received” by the offspring, with distinct levels for each unordered parental pair, and with correlation
between effects received by half-sibs (one shared parent).
corresponds to var(GCA)/[2*var(GCA)+var(SCA)] and is necessarily
.
See the X.GCA
documentation for similar constructs for fixed effects.
The functions return corrFamily descriptors whose general format is described in corrFamily
. The ones produced by ranGCA
and antisym
are atypical in that only their Af
element is non-trivial.
Lynch, M., Walsh, B. (1998) Genetics and analysis of quantitative traits. Sinauer, Sunderland, Mass.
#### Simulate dyadic data set.seed(123) nind <- 10 # Beware data grow as O(nind^2) x <- runif(nind^2) id12 <- expand.grid(id1=seq(nind),id2=seq(nind)) id1 <- id12$id1 id2 <- id12$id2 u <- rnorm(nind,mean = 0, sd=0.5) ## additive individual effects: y <- 0.1 + 1*x + u[id1] + u[id2] + rnorm(nind^2,sd=0.2) ## Same with non-additive individual effects: dist.u <- abs(u[id1] - u[id2]) z <- 0.1 + 1*x + dist.u + rnorm(nind^2,sd=0.2) ## anti-smmetric individual effects: t <- 0.1 + 1*x + u[id1] - u[id2] + rnorm(nind^2,sd=0.2) dyaddf <- data.frame(x=x, y=y, z=z, t=t, id1=id1,id2=id2) # : note that this contains two rows per dyad, which avoids identifiability issues. # Enforce that interactions are between distinct individuals (not essential for the fit): dyaddf <- dyaddf[- seq.int(1L,nind^2,nind+1L),] # Fits: (addfit <- fitme(y ~x +ranGCA(1|id1+id2), data=dyaddf)) # # practically equivalent to: # (fitme(y ~x +diallel(1|id1+id2, fixed=0.49999), data=dyaddf)) (antifit <- fitme(t ~x +antisym(1|id1+id2), data=dyaddf)) (distfit <- fitme(z ~x +diallel(1|id1+id2), data=dyaddf))
#### Simulate dyadic data set.seed(123) nind <- 10 # Beware data grow as O(nind^2) x <- runif(nind^2) id12 <- expand.grid(id1=seq(nind),id2=seq(nind)) id1 <- id12$id1 id2 <- id12$id2 u <- rnorm(nind,mean = 0, sd=0.5) ## additive individual effects: y <- 0.1 + 1*x + u[id1] + u[id2] + rnorm(nind^2,sd=0.2) ## Same with non-additive individual effects: dist.u <- abs(u[id1] - u[id2]) z <- 0.1 + 1*x + dist.u + rnorm(nind^2,sd=0.2) ## anti-smmetric individual effects: t <- 0.1 + 1*x + u[id1] - u[id2] + rnorm(nind^2,sd=0.2) dyaddf <- data.frame(x=x, y=y, z=z, t=t, id1=id1,id2=id2) # : note that this contains two rows per dyad, which avoids identifiability issues. # Enforce that interactions are between distinct individuals (not essential for the fit): dyaddf <- dyaddf[- seq.int(1L,nind^2,nind+1L),] # Fits: (addfit <- fitme(y ~x +ranGCA(1|id1+id2), data=dyaddf)) # # practically equivalent to: # (fitme(y ~x +diallel(1|id1+id2, fixed=0.49999), data=dyaddf)) (antifit <- fitme(t ~x +antisym(1|id1+id2), data=dyaddf)) (distfit <- fitme(z ~x +diallel(1|id1+id2), data=dyaddf))
This experimental function displays information about parameter values for which some numerical problems have occurred. Some warnings suggest its use.
Numerical problems may occur if the information matrix (for the augmented linear model used in the iteratively reweighted least-squares algorithm) is nearly singular. spaMM may try to check whether such singularity occurs when this algorithm has not converged. But itself may be slow so it is not performed systematically for large matrices. spaMM.options(diagnose_conv=<integer>)
may be used to control the maximum size of matrices for which the check is performed.
When “outer” generic optimization is performed, information is reported about the range of parameter values for which problems occurred, (see Value). The fit object divinfo
element may also contain more informative tables of parameter values. This information is currently missing for “inner”-optimized parameters.
div_info(object, ...)
div_info(object, ...)
object |
An object of class |
... |
Currently not used |
Used mainly for the side effects (printed output) but returns invisibly either a single parameter vector (if a single numerical problem occurred) or a matrix of parameter ranges, or NULL if there is no problem to report.
# Tragically ;-), no simple example of numerical problems # that can be diagnosed by div_info() is currently available.
# Tragically ;-), no simple example of numerical problems # that can be diagnosed by div_info() is currently available.
This extracts the number of degrees of freedom for a model, in the usual sense for likelihood-ratio tests: a count of number of fitted parameters, distinguishing different classes of parameters (see Value).
DoF(object)
DoF(object)
object |
A fitted-model object, of class |
The output distinguishes counts of random-effect vs residual-dispersion parameters, following the conceptual distinction between effects that induce correlations between different levels of the resonse vs. observation-level effects. However, a residual-dispersion component can be declared as a random effect, so that the counts for logically equivalent models may differ according to the way a model was declared. For example if residual dispersion for an LLM is declared as an observation-level random effect while phi
is fixed, the p_lambda
component will include 1 df for what would otherwise be accounted by the p_rdisp
component. A more involved case where the same contrast happens is when a negative-binomial model (with a residual-dispersion shape
parameter) is declared as a Poisson-gamma mixture model (with a varaince parameter for the Gamma-distributed individual-level random effect).
A vector with possible elements p_fixef
, p_lambda
, p_corrPars
and p_rdisp
for, respectively, the number of fixed-effect coefficients of the main-response model, the number of random-effect variance parameters, the number of random-effect correlation parameters, and the number of residual dispersion parameters (the latter being itself, for a mixed-effect residual-dispersion model, the sum of such components).
df.residual.HLfit
; get_any_IC
for extracting effective degrees of freedom considered in the model-selection literature; as_LMLT
for access to the effective degrees of freedom considered in Satterthwaite's test and its extentions.
interface to apply some function fn
in parallel on columns of a matrix. It is not logically restricted to mixed-effect applications, hence it can be used more widely. Depending on the nb_cores
argument, parallel or serial computation is performed, calling the future.apply::future_apply
function. A socket cluster is used by default for parallel computations, but a fork cluster can be requested on linux and alike operating systems by using argument cluster_args=list(type="FORK")
.
dofuture(newresp, fn, nb_cores=NULL, fit_env, control=list(), cluster_args=NULL, debug.=FALSE, iseed=NULL, showpbar="ignored", pretest_cores="ignored", ... )
dofuture(newresp, fn, nb_cores=NULL, fit_env, control=list(), cluster_args=NULL, debug.=FALSE, iseed=NULL, showpbar="ignored", pretest_cores="ignored", ... )
newresp |
A matrix on whose columns |
fn |
Function whose first argument is named |
nb_cores |
Integer. Number of cores to use for parallel computations. If >1, a cluster of |
fit_env |
(for socket clusters only:) An environment, or a list, containing variables to be exported on the nodes of the cluster (by |
control |
A list. The only effective control is |
cluster_args |
A list of arguments passed to |
debug. |
(for socket clusters only:) For debugging purposes. Effect, if any, is to be defined by the |
iseed |
Integer, or NULL. If an integer, it is used to initialize |
showpbar , pretest_cores
|
Currently ignored; for consistency with |
... |
Further arguments to be passed (unevaluated) to |
The result of calling future.apply
. If the progressr
package is loaded, a side-effect of dofuture
is to show a progress bar with character 'S' or 'P' or 'F' depending on parallelisation status (serial/socket/fork).
dopar
for an alternative implementation of (essentially) the same functionalities, and wrap_parallel
for its differences from dofuture
.
## Not run: if (requireNamespace("future.apply", quietly = TRUE)) { # Useless function, but requiring some argument beyond the first foo <- function(y, somearg, ...) { if ( is.null(somearg) || TRUE ) length(y) } # Whether FORK can be used depends on OS and whether Rstudio is used: dofuture(matrix(1,ncol=4,nrow=3), foo, fit_env=list(), somearg=NULL, nb_cores=2, cluster_args=list(type="FORK")) } ## End(Not run)
## Not run: if (requireNamespace("future.apply", quietly = TRUE)) { # Useless function, but requiring some argument beyond the first foo <- function(y, somearg, ...) { if ( is.null(somearg) || TRUE ) length(y) } # Whether FORK can be used depends on OS and whether Rstudio is used: dofuture(matrix(1,ncol=4,nrow=3), foo, fit_env=list(), somearg=NULL, nb_cores=2, cluster_args=list(type="FORK")) } ## End(Not run)
dopar
and combinepar
are interfaces primarily designed to apply some function fn
in parallel on columns of a matrix, although other uses are possible. Depending on the nb_cores
argument, parallel or serial computation is performed. A socket cluster is used by default for parallel computations, but a fork cluster can be requested on linux and alike operating systems by using argument cluster_args=list(type="FORK")
.
dopar
has been designed to provide by default a progress bar in all evaluations contexts. A drawback is that different procedures are called depending e.g. on the type of cluster, with different possible controls. In particular, foreach
is called in some cases but not others, so non-trivial values of its .combine
control are not always enforced. The alternative interface combinepar
will always use foreach
, and will still try to provide by default a progress bar but may fail to do so in some cases (see Details).
dopar(newresp, fn, nb_cores = NULL, fit_env, control = list(.final=function(v) if( ! is.list(v[[1]])) {do.call(cbind,v)} else v), cluster_args = NULL, debug. = FALSE, iseed = NULL, showpbar = eval(spaMM.getOption("barstyle")), pretest_cores =NULL, ...) combinepar(newresp, fn, nb_cores = NULL, cluster=NULL, fit_env, control = list(.final=function(v) if( ! is.list(v[[1]])) {do.call(cbind,v)} else v), cluster_args = NULL, debug. = FALSE, iseed = NULL, showpbar = eval(spaMM.getOption("barstyle")), pretest_cores =NULL, ...)
dopar(newresp, fn, nb_cores = NULL, fit_env, control = list(.final=function(v) if( ! is.list(v[[1]])) {do.call(cbind,v)} else v), cluster_args = NULL, debug. = FALSE, iseed = NULL, showpbar = eval(spaMM.getOption("barstyle")), pretest_cores =NULL, ...) combinepar(newresp, fn, nb_cores = NULL, cluster=NULL, fit_env, control = list(.final=function(v) if( ! is.list(v[[1]])) {do.call(cbind,v)} else v), cluster_args = NULL, debug. = FALSE, iseed = NULL, showpbar = eval(spaMM.getOption("barstyle")), pretest_cores =NULL, ...)
newresp |
A matrix on whose columns |
fn |
Function whose first argument is named |
nb_cores |
Integer. Number of cores to use for parallel computations. If >1 (and no cluster is provided by the |
cluster |
(for |
fit_env |
(for socket clusters only:) An environment, or a list, containing variables to be exported on the nodes of the cluster (by |
control |
A list following the
|
cluster_args |
A list of arguments passed to |
debug. |
(for socket clusters only:) For debugging purposes. Effect, if any, is to be defined by the |
iseed |
(all parallel contexts:) Integer, or NULL. If an integer, it is used as the |
showpbar |
(for socket clusters only:) Controls display of progress bar. See |
pretest_cores |
(for socket clusters only:) A function to run on the cores before running |
... |
Further arguments to be passed (unevaluated) to |
Control of random numbers through the "L'Ecuyer-CMRG"
generator and the iseed
argument is not sufficient for consistent results when the doSNOW
parallel backend is used, so if you really need such control in a fn
using random numbers, do not use doSNOW
. Yet, it is fine to use doSNOW
for bootstrap procedures in spaMM, because the fitting functions do not use random numbers: only sample simulation uses them, and it is not performed in parallel.
combinepar
calls foreach::%dopar%
which assumes that a cluster has been declared using a suitable backend such as doSNOW
, doFuture
or doParallel
. If only the latter is available, no progress bar is displayed. A method to render a bar when doParallel
is used can be found on the Web, but that bar is not a valid progress bar as it is displayed only after all the processes have been run.
The result of calling foreach
, pbapply
or mclapply
, as dependent on the control
argument and the interface used. A side-effect of either interface is to show a progress bar whose character informs about the type of parallelisation performed: a "F"
or default "="
character for fork clusters, a "P"
for parallel computation via foreach
and doSNOW
, a "p"
for parallel computation via foreach
and doFuture
or via pbapply
, and "s"
for serial computation foreach
and doParallel
or via pbapply
.
dofuture
is yet another interface with (essentially) the same functionalities as dopar
. See the documentation of the wrap_parallel
option for its differences from dopar
.
## See source code of spaMM_boot() ## Not run: # Useless function, but requiring some argument beyond the first foo <- function(y, somearg, ...) { if ( is.null(somearg) || TRUE ) length(y) } # Whether FORK can be used depends on OS and whether Rstudio is used: dopar(matrix(1,ncol=4,nrow=3), foo, fit_env=list(), somearg=NULL, nb_cores=2, cluster_args=list(type="FORK")) ## End(Not run)
## See source code of spaMM_boot() ## Not run: # Useless function, but requiring some argument beyond the first foo <- function(y, somearg, ...) { if ( is.null(somearg) || TRUE ) length(y) } # Whether FORK can be used depends on OS and whether Rstudio is used: dopar(matrix(1,ncol=4,nrow=3), foo, fit_env=list(), somearg=NULL, nb_cores=2, cluster_args=list(type="FORK")) ## End(Not run)
Drop single terms from the model. The drop1
method for spaMM fit objects is conceived to replicate the functionality, output format and details of pre-existing methods for similar models. Results for LMs and GLMs should replicate base R drop1
results, with some exceptions:
* somewhat stricter default check of non-default scope
argument;
* Because the dispersion estimates for Gamma GLMs differ between stats::glm
and spaMM fits (see Details in method
), some tests may differ too; results from spaMM REML fits being closer than ML fits to those from glm()
fits;
* AIC values reported in tables are always the marginal AIC as computed by AIC.HLfit
, while drop1.glm
may report confusing (to me, at least) values (see AIC.HLfit
) for reasons that seem to go beyond differences in dispersion estimates.
For LMMs, ANOVA tables are provided by interfacing lmerTest::anova
(with non-default type
).
For other classes of models, a table of likelihood ratio tests is returned, each test resulting from a call to LRT
.
## S3 method for class 'HLfit' drop1(object, scope, method="", check_marg=NULL, check_time=60, ...)
## S3 method for class 'HLfit' drop1(object, scope, method="", check_marg=NULL, check_time=60, ...)
object |
Fit object returned by a spaMM fitting function. |
scope |
Default “scope” (terms to be tested, specified as a formula, see Examples) is determined by applying |
method |
Only non-default value is |
check_marg |
NULL or boolean: whether effects should be checked for marginality. By default, this check is performed when a non-default scope is specified, and then no test is reported for terms that do not satisfy the marginality condition. If |
check_time |
numeric: whether to output some information when the execution time of |
... |
Further arguments passed from or to methods, or to |
As for the ANOVA-table functionality, it has been included here mainly to provide access to F tests (including, for LMMs, the “Satterthwaite method”, using pre-existing procedures as template or backend for expediency and familiarity. The procedures for specific classes of models have various limitations, e.g., none of them handle models with variable dispersion parameter. For classes of models not well handled by these procedures (by design or due to the experimental nature of the recent implementation), method="LRT"
can still be applied (and will be applied by default for GLMMs).
The return format is that of the function called (lmerTest::drop1
for LMMs), or emulated (base drop1
methods for LMs or GLMs), or is a data frame whose rows are each the result of calling LRT
.
as_LMLT
for the interface to lmerTest::drop1
.
data("wafers") #### GLM wfit <- fitme(y ~ X1+X2+X1*X3+X2*X3+I(X2^2), family=Gamma(log), data=wafers) drop1(wfit, test = "F") drop1(wfit, test = "F", scope= ~ X1 + X1 * X3 ) # note the message! #### LMM if(requireNamespace("lmerTest", quietly=TRUE)) { lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) drop1(lmmfit) # => Satterthwaite method here giving p-values quite close to # traditional t-tests given by: summary(lmmfit, details=list(p_value=TRUE)) } #### GLMM wfit <- fitme(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), rand.family=inverse.Gamma(log), resid.model = ~ X3+I(X3^2) , data=wafers) drop1(wfit) drop1(wfit, scope= ~ X1 + X1 * X3 ) # note the message!
data("wafers") #### GLM wfit <- fitme(y ~ X1+X2+X1*X3+X2*X3+I(X2^2), family=Gamma(log), data=wafers) drop1(wfit, test = "F") drop1(wfit, test = "F", scope= ~ X1 + X1 * X3 ) # note the message! #### LMM if(requireNamespace("lmerTest", quietly=TRUE)) { lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) drop1(lmmfit) # => Satterthwaite method here giving p-values quite close to # traditional t-tests given by: summary(lmmfit, details=list(p_value=TRUE)) } #### GLMM wfit <- fitme(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), rand.family=inverse.Gamma(log), resid.model = ~ X3+I(X3^2) , data=wafers) drop1(wfit) drop1(wfit, scope= ~ X1 + X1 * X3 ) # note the message!
eval_replicate
is the default simuland
function applied to simulated bootstrap samples by likelihood-ratio testing functions (fixedLRT, LRT, anove.HLfit
). This documentation presents the requirements and possible features of this function and of possible user-defined alternatives.
An alternative function spaMM:::.eval_replicate2
is also provided. It is slower, as it refits the models compared with different initial values for random-effect parameters, which is useful in some difficult cases where initial values matter. The eval_replicate
function may also refit the “full” models with different initial values when the logLik of the refitted full model is substantially lower than that of the refitted null model. “Substantially” means that a tolerance of 1e-04
is applied to account for inaccuracies of numerical maximization.
eval_replicate(y)
eval_replicate(y)
y |
a response vector on which a previously fitted model may be refitted. |
likelihood-ratio testing functions have a debug.
argument whose effect depends on the simuland
function. The default behaviour is thus defined by eval_replicate
, as: if debug.=TRUE
, upon error in the fitting procedures, dump.frames
will be called, in which case a dump file will be written on disk; and a list with debugging information will be returned (so that, say, pbapply
will not return a matrix). This behaviour may change in later versions, so non-default debug.
values should not be used in reproducible code. In serial computation, debug.=2
may induce a stop
; this should not happen in parallel computation because the calling functions check against debug.==2
.
Essential information such as the originally fitted models is passed to the function not as arguments but through its environment, which is controlled by the calling functions (see the eval_replicate
source code to know which are these arguments). Users should thus not assume that they can control their own simuland
function's environment as this environment will be altered.
Advanced users can define their own simuland
function. The eval_replicate
source code provides a template showing how to use the function's environment. The Example below illustrates another approach augmenting eval_replicate
. A further example is provided in the filetests/testthat/test-LRT-boot.R
, using ... to pass additional arguments beyond response values.
A vector of the form c(full=logLik(
<refitted full model>),null=logLik(
<refitted null model>)
; or possibly in debugging contexts, a list with the same elements each with some additional information provided as attribute.
Calling functions fixedLRT, LRT
.
## Not run: # Simple wrapper enhancing the default 'simuland' # with a call to some obscure option, and dealing with # the need to pass the environment assigned to 'simuland' eval_with_opt <- function(y) { spaMM.options(some_obscure_option="some_obscure_value") eval_rep <- spaMM:::.eval_replicate environment(eval_rep) <- parent.env(environment()) # passing the environment eval_rep(y) } ## End(Not run)
## Not run: # Simple wrapper enhancing the default 'simuland' # with a call to some obscure option, and dealing with # the need to pass the environment assigned to 'simuland' eval_with_opt <- function(y) { spaMM.options(some_obscure_option="some_obscure_value") eval_rep <- spaMM:::.eval_replicate environment(eval_rep) <- parent.env(environment()) # passing the environment eval_rep(y) } ## End(Not run)
spaMM is conceived to minimize installation issues but it nevertheless suggests using some external libraries. These are all accessed through R packages so their installation should be easy when installing binary packages. The Details below give hints for installing packages from source. They may all be obsolete if you are using the Rtools42 on Windows. For all cases not considered below, help yourself. If you are using the Rtools40 on Windows, you should have a look at the package manager in the Rtools40 bash shell.
The ROI.plugin.glpk
package requires the Rglpk
package, which itself requires the external glpk
library. For the latter, Debian-ists and alikes should sudo apt-get install libglpk-dev
. MacOSX users should brew install glpk
if using brew
; Windows users should try usingpacman -S mingw-w64-x86_64-glpk
in the Rtools40 bash shell, together withSys.setenv(GLPK_HOME = "$(MINGW_PREFIX)")
in the R session (but I have not fully tested this; previously I had to install glpk from https://sourceforge.net/projects/winglpk/).
The nloptr
package requires the external NLopt
library. Windows users should try usingpacman -S mingw-w64-x86_64-nlopt
in the Rtools40 bash shell (but again I have not fully tested this; see also the README of nloptr
).
To install nloptr 2.0.0 from sources on Debian one may have to install libnlopt-dev: sudo apt-get install libnlopt-dev
Likewise for gmp one may need to install libgmp3-dev
.
Most extractors are methods of generic functions defined in base R (see Usage), for which the base documentation may be useful.
formula
extracts the model formula.family
extracts the response family.terms
extracts the formula, with attributes describing the fixed-effect terms.nobs
returns the length of the response vector.logLik
extracts the log-likelihood (exact or approximated).dev_resids
returns a vector of squared (unscaled) deviance residuals (the summands defined for GLMs in McCullagh and Nelder 1989, p. 34; see Details of LL-family
for other response families. deviance
returns the sum of squares of these deviance residuals, possibly weighted by prior weights (consistently with stats::deviance
. See residuals.HLfit
for details and comparison with related extractors.fitted
extracts fitted values.response
extracts the response (as a vector).fixef
extracts the fixed effects coefficients, .
ranef
extracts the predicted random effects, Lv (default since version 1.12.0), or optionally u (see random-effects
for definitions). print.ranef
controls their printing. getDistMat
returns a distance matrix for a geostatistical (Matérn etc.) random effect.df.residual
extracts residual degrees-of-freedom for fitted models (here number of observations minus number of parameters of the model except residual dispersion parameters).
wweights
extracts prior weights (as defined by the fitting functions's prior.weights
argument).
## S3 method for class 'HLfit' formula(x, which="hyper", ...) ## S3 method for class 'HLfit' family(object, ...) ## S3 method for class 'HLfit' terms(x, ...) ## S3 method for class 'HLfit' nobs(object, ...) ## S3 method for class 'HLfit' logLik(object, which, ...) ## S3 method for class 'HLfit' fitted(object, ...) ## S3 method for class 'HLfit' fixef(object, na.rm=NULL, ...) ## S3 method for class 'HLfit' ranef(object, type = "correlated", ...) ## S3 method for class 'ranef' print(x, max.print = 40L, ...) ## S3 method for class 'HLfit' deviance(object, ...) ## S3 method for class 'HLfit' df.residual(object, ...) ## S3 method for class 'HLfit' weights(object, type, ...) ## getDistMat(object, scaled=FALSE, which = 1L) response(object,...) dev_resids(object,...)
## S3 method for class 'HLfit' formula(x, which="hyper", ...) ## S3 method for class 'HLfit' family(object, ...) ## S3 method for class 'HLfit' terms(x, ...) ## S3 method for class 'HLfit' nobs(object, ...) ## S3 method for class 'HLfit' logLik(object, which, ...) ## S3 method for class 'HLfit' fitted(object, ...) ## S3 method for class 'HLfit' fixef(object, na.rm=NULL, ...) ## S3 method for class 'HLfit' ranef(object, type = "correlated", ...) ## S3 method for class 'ranef' print(x, max.print = 40L, ...) ## S3 method for class 'HLfit' deviance(object, ...) ## S3 method for class 'HLfit' df.residual(object, ...) ## S3 method for class 'HLfit' weights(object, type, ...) ## getDistMat(object, scaled=FALSE, which = 1L) response(object,...) dev_resids(object,...)
object |
An object of class |
type |
For |
which |
* For |
na.rm |
Whether to include NA values for missing coefficients of rank-deficient model matrices. Default is to exclude them for mixed models and to include them for other ones. See Details for the underlying reason. |
scaled |
If |
x |
For |
max.print |
Controls |
... |
Other arguments that may be needed by some method. |
For rank-deficient model matrices, base R procedures lm
and glm
estimate coefficients for a rank-trimmed matrix and coefficient()
returns a full-length vector completed with NA values for coefficients not estimated, while the lme4 fixef
method returns a trimmed vector. spaMM has long followed the base R convention for all models but this may impede use of some post-fit procedures initially conceived for lme4 objects (such as lmerTest procedures for LMMs). So now fixef.HLfit
trims the vector by default for mixed-effect models only. The default is thus to maximize consistency/compatibility with preexisting procedures despite their inconsistencies with each other.
formula
returns a formula
, except a list of them from fitmv()
output.
terms
returns an object of class c("terms", "formula")
which contains the terms representation of a symbolic model. See terms.object
for its structure. terms(<fitmv() result>)
returns a list of such terms.
Other return values are numeric (for logLik
), vectors (most cases), matrices or dist objects (for getDistMat
), or a family object (for family
). ranef
returns a list of vectors or matrices (the latter for random-coefficient terms).
McCullagh, P. and Nelder J. A. (1989) Generalized linear models. Second ed. Chapman & Hall: London.
See summary.HLfit
whose return value include the tables of fixed-effects coefficients and random-effect variances displayed by the summary, residuals.HLfit
to extract various residuals, residVar
to extract residual variances or information about residual variance models, hatvalues
to extract leverages, get_matrix
to extract the model matrix and derived matrices, and vcov.HLfit
to extract covariances matrices from a fit, get_RLRsim_args
to extract arguments for (notably) tests of random effects in LMMs.
data("wafers") m1 <- fitme(y ~ X1+X2+(1|batch), data=wafers) fixef(m1) ranef(m1) data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) getDistMat(fitobject)
data("wafers") m1 <- fitme(y ~ X1+X2+(1|batch), data=wafers) fixef(m1) ranef(m1) data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) getDistMat(fitobject)
regularize
can be used to regularize (nearly-)singular correlation matrices. It may also be used to regularize covariance matrices but will not keep their diagonal constant. Use on other types of matrices may give nonsense. The regularization corrects the diagonal of matrices with high condition number so that the condition number of a corrected matrix is the maximum value specified by maxcondnum
. For that purpose, it needs the extreme eigenvalues of the matrix, by default provided by the function extreme_eig
. Calls functions from RSpectra if available, and falls back on base functions otherwise.
extreme_eig(M, symmetric, required = TRUE) regularize(A, EEV=extreme_eig(A,symmetric=TRUE), maxcondnum=1e12)
extreme_eig(M, symmetric, required = TRUE) regularize(A, EEV=extreme_eig(A,symmetric=TRUE), maxcondnum=1e12)
M |
Square matrix. Sparse matrices of class |
A |
Square matrix as |
symmetric |
Whether the matrix is symmetric. Helpful to select efficient methods for this case if the matrix class does not implies its symmetry. |
required |
Whether the computation should be attempted independently of the size of the matrix. |
EEV |
Two extreme eigenvalue in the return format of |
maxcondnum |
Target condition number when regularization is performed |
extreme_eig
returns a vector of length 2, the largest and the smallest eigenvalues in this order.
regularize
returns a matrix, possibly in sparse format.
H10 <- Matrix::Hilbert(10) extreme_eig(H10,symmetric=TRUE) # ratio > 1e13 rH10 <- regularize(H10) extreme_eig(rH10,symmetric=TRUE) # ratio = 1e12
H10 <- Matrix::Hilbert(10) extreme_eig(H10,symmetric=TRUE) # ratio > 1e13 rH10 <- regularize(H10) extreme_eig(rH10,symmetric=TRUE) # ratio = 1e12
This is a common interface for fitting most models that spaMM can fit, from linear models to mixed models with non-gaussian random effects, therefore substituting to corrHLfit
, HLCor
and HLfit
. By default, it uses ML rather than REML (differing in this respect from the other fitting functions). It may use “outer optimization”, i.e., generic optimization methods for estimating all dispersion parameters, rather than the iterative methods implemented in HLfit
. The results of REML fits of non-gaussian mixed models by these different methods may (generally slightly) differ. Outer optimization should generally be faster than the alternative algorithms for large data sets when the residual variance model is a single constant term (no structured dispersion). For mixed models, fitme
by default tries to select the fastest method when both can be applied, but precise decision criteria are subject to change in the future. corrHLfit
(with non-default arguments to control the optimization method most suitable to a particular problem) may be used to ensure better consistency over successive versions of spaMM
.
fitme(formula, data, family = gaussian(), init = list(), fixed = list(), lower = list(), upper = list(), resid.model = ~1, init.HLfit = list(), control = list(), control.dist = list(), method = "ML", HLmethod = method, processed = NULL, nb_cores = NULL, objective = NULL, weights.form = NULL, ...)
fitme(formula, data, family = gaussian(), init = list(), fixed = list(), lower = list(), upper = list(), resid.model = ~1, init.HLfit = list(), control = list(), control.dist = list(), method = "ML", HLmethod = method, processed = NULL, nb_cores = NULL, objective = NULL, weights.form = NULL, ...)
formula |
Either a linear model |
data |
A data frame containing the variables in the response and the model formula. |
family |
|
init |
An optional list of initial values for correlation and/or dispersion parameters and/or response family parameters, e.g.
|
fixed |
A list similar to |
lower |
An optional (sub)list of values of the parameters specified through |
upper |
Same as |
resid.model |
See identically named |
init.HLfit |
See identically named |
control.dist |
See |
method , HLmethod
|
Character: the fitting method to be used, such as |
weights.form |
Specification of prior weights by a one-sided formula: use |
control |
A list of (rarely needed) control parameters, with possible elements:
|
nb_cores |
For development purpose, not documented. |
processed |
For programming purpose, not documented. |
objective |
For development purpose, not documented. |
... |
Optional arguments passed to (or operating as if passed to) |
For approximations of likelihood, see method
. For the possible structures of random effects, see random-effects
,
For phi
, lambda
, and ranCoefs
, fitme
may or may not use the internal fitting methods of HLfit
. The latter methods are well suited for structured dispersion models, but require computations which can be slow for large datasets. Therefore, fitme
tends to outer-optimize by default for large datasets, unless there is a non-trivial resid.model
. The precise criteria for selection of default method by fitme
are liable to future changes.
Further, the internal fitting methods of HLfit
also provide some more information such as the “cond. SE” (about which see warning in Details of HLfit
). To force the evaluation of such information after an outer-optimization by a fitme
call, use the control$refit
argument (see Example). Alternatively (and possibly of limited use), one can force inner-optimization of lambda
for a given random effect, or of phi
, by setting it to NaN
in init
(see Example using ‘blackcap’ data). The same syntax may be tried for phi
.
The return value of an HLCor
or an HLfit
call, with additional attributes. The HLCor
call is evaluated at the estimated correlation parameter values. These values are included in the return object as its $corrPars
member. The attributes added by fitme
include the original call of the function (which can be retrived by getCall
(<fitted object>), and information about the optimization call within fitme
.
## Examples with Matern correlations ## A likelihood ratio test based on the ML fits of a full and of a null model. data("blackcap") (fullfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) ) (nullfit <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap)) ## p-value: 1-pchisq(2*(logLik(fullfit)-logLik(nullfit)),df=1) ## See ?spaMM for examples of conditional autoregressive model and of non-spatial models. ## Contrasting different optimization methods: # We simulate Gamma deviates with mean mu=3 and variance=2, # ie. phi= var/mu^2= 2/9 in the (mu, phi) parametrization of a Gamma # GLM; and shape=9/2, scale=2/3 in the parametrisation of rgamma(). # Note that phi is not equivalent to scale: # shape = 1/phi and scale = mu*phi. set.seed(123) gr <- data.frame(y=rgamma(100,shape=9/2,scale=2/3)) # Here fitme uses HLfit methods which provide cond. SE for phi by default: fitme(y~1,data=gr,family=Gamma(log)) # To force outer optimization of phi, use the init argument: fitme(y~1,data=gr,family=Gamma(log),init=list(phi=1)) # To obtain cond. SE for phi after outer optimization, use the 'refit' control: fitme(y~1,data=gr,family=Gamma(log),,init=list(phi=1), control=list(refit=list(phi=TRUE))) ## or ...refit=TRUE... ## Outer-optimization is not necessarily the best way to find a global maximum, # particularly when there is little statistical information in the data: if (spaMM.getOption("example_maxtime")>1.6) { data("blackcap") fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) # poor # Compare with the following two ways of avoiding outer-optimization of lambda: corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML") fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=list(lambda=NaN)) } ## see help("COMPoisson"), help("negbin"), help("Loaloa"), etc., for further examples.
## Examples with Matern correlations ## A likelihood ratio test based on the ML fits of a full and of a null model. data("blackcap") (fullfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) ) (nullfit <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap)) ## p-value: 1-pchisq(2*(logLik(fullfit)-logLik(nullfit)),df=1) ## See ?spaMM for examples of conditional autoregressive model and of non-spatial models. ## Contrasting different optimization methods: # We simulate Gamma deviates with mean mu=3 and variance=2, # ie. phi= var/mu^2= 2/9 in the (mu, phi) parametrization of a Gamma # GLM; and shape=9/2, scale=2/3 in the parametrisation of rgamma(). # Note that phi is not equivalent to scale: # shape = 1/phi and scale = mu*phi. set.seed(123) gr <- data.frame(y=rgamma(100,shape=9/2,scale=2/3)) # Here fitme uses HLfit methods which provide cond. SE for phi by default: fitme(y~1,data=gr,family=Gamma(log)) # To force outer optimization of phi, use the init argument: fitme(y~1,data=gr,family=Gamma(log),init=list(phi=1)) # To obtain cond. SE for phi after outer optimization, use the 'refit' control: fitme(y~1,data=gr,family=Gamma(log),,init=list(phi=1), control=list(refit=list(phi=TRUE))) ## or ...refit=TRUE... ## Outer-optimization is not necessarily the best way to find a global maximum, # particularly when there is little statistical information in the data: if (spaMM.getOption("example_maxtime")>1.6) { data("blackcap") fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) # poor # Compare with the following two ways of avoiding outer-optimization of lambda: corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML") fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=list(lambda=NaN)) } ## see help("COMPoisson"), help("negbin"), help("Loaloa"), etc., for further examples.
This function extends the fitme
function to fit a joint model for different responses (following possibly different response families) sharing some random-effects, including a new type of random effect defined to exhibit correlations across different responses (see mv
).
It is also possible to declare shared fixed-effect coefficients among different submodels, using the X2X
argument.
Only a few features available for analysis of univariate response may not yet work (see Details).
fitmv(submodels, data, fixed=NULL, init=list(), lower=list(), upper=list(), control=list(), control.dist = list(), method="ML", init.HLfit=list(), X2X=NULL, ...)
fitmv(submodels, data, fixed=NULL, init=list(), lower=list(), upper=list(), control=list(), control.dist = list(), method="ML", init.HLfit=list(), X2X=NULL, ...)
submodels |
A list of sublists each specifying a model for each univariate response. The names given to each submodel in the main list are currently ignored. The names and syntax of elements within each sublist are those of a
|
data |
A data frame containing the variables in the response and the model formulas. |
fixed |
A list of fixed values of the parameters controlling random effects. The syntax is that of the same argument in |
init , lower , upper
|
Lists of initial values or bounds. The syntax is that of the same arguments in |
control |
A list of control parameters, with possible elements as described for |
control.dist |
See |
method |
Character: the fitting method to be used, such as |
init.HLfit |
See identically named |
X2X |
NULL, or a matrix M by which one can specify, as |
... |
Optional arguments passed to (or operating as if passed to) |
Matching random effects across submodels, and referring to them;
Random effects are recognized as identical across submodels by matching the formula terms. As shown in the Examples, if the two models formulas share the (1|clinic)
term, this term is recognized as a single random effect shared between the two responses. But the (1|clinic)
and (+1|clinic)
terms are recognized as distinct random effects. In that case, the init
argument init=list(lambda=c('1'=1,'2'=0.5))
is shown to refer to these by names 1,2
... where the order is defined as the order of first appearance of the terms across the model formulas in the order of the submodels
list.
Alternatively, the syntax fixed=list(lambda=c('clinic.1'=0.5,'clinic'=1))
works: this syntax makes order of input irrelevant but assumes that the user guesses names correctly (these are typically the names that appear in the summary of lambda values from the fit object or, more programmatically, names(<fit object>$lambda.object$print_namesTerms)
). Finally, fixed values of parameters can also be specified through each sub-model, with indices referring to the order of random effects with each model.
The matching of random-effect terms occurs after expansion of multIMRF
terms, if any. This may have subtle consequences if two multIMRF terms differ only by their number of levels, as some of the expanded IMRF terms are then shared.
Capacities and limitations:
Practically all features of models that can be fitted by fitme
should be available: this includes all combinations of GLM response families, residual dispersion models, and all types of random-effect terms, whether autocorrelated or not. Among the arguments handled through the ..., covStruct
, distMatrix
, corrMatrix
should be effective; control.HLfit$LevenbergM
and verbose=c(TRACE=TRUE)
will work but some other controls available in fitme
may not. Usage of the REMLformula
argument is restricted as it cannot be used to specify a non-standard REML correction (but the more useful keepInREML
attribute for fixed fixed-effect coefficients is handled).
The multi
family-like syntax for multinomial models should not be used, but fitmv
could provide other means to model multinomial responses.
Most post-fit functions work, at least with default arguments. This includes point predict
ion and prediction variances calculations sensu lato, including with newdata
; but also simulate
, spaMM_boot
, confint
, anova
, update_resp
, and update
. The re.form
argument now works for predict
and simulate
. Bootstrap computation may require special care for models where simulation of one response variable may depend on draws of another one (see Hurdle model example in the “Gentle introdution” to spaMM,
https://gitlab.mbb.univ-montp2.fr/francois/spamm-ref/-/blob/master/vignettePlus/spaMMintro.pdf).
Prediction functions try to handle most forms of missing information in newdata
(including information missing for a residual-dispersion model when predcitiosn fro mit are needed: see Examples). As information may be missing for some submodels but not others, different numbers of predictions are then returned for different submodels. As for univariate-response models, predict
will return point predictions as a single 1-column matrix, here concatenating the prediction results of the different submodels. The nobs
attribute specifies how may values pertain to each submodel.
Some plotting functions may fail. update.formula
fails (see update_formulas
for details). terms
returns a list, which is not usable by other base R functions. stats::step
is a good example of resulting limitations, as it is currently unable to perform any sensible operation on fitmv
output. spaMM::MSFDR
which calls stats::step
likewise fails.
multcomp::glht
fails.
A perhaps not entirely satisfying feature is that simulate
by default stacks the results of simulating each submodel in a single vector. Some non-trivial reformatting may then be required to include such simulation results in a suitable newdata
data frame with (say) sufficient information for prediction of all responses. The syntax update_resp(<fit>, newresp = simulate(<fit>, ...), evaluate = FALSE)$data
may be particularly useful to reformat simulation results in this perspective.
Which arguments belong to submodels
?:
Overall, arguments specifying individual submodels should go into submodels
, while other arguments of fitmv
should be those potentially affecting several submodels (notably, random-effect structures, lower
, and upper
) and fitting controls (such as init
and init.HLfit
). One rarely-used exception is REMLformula
which controls the fitting method but should be specified through the submodels
.
The function proceeds by first preprocessing all submodels independently, before merging the resulting information by matching random effects across submodels. The merging operation includes some checks of consistency across submodels, implying that redundant arguments may be needed across submodels (e.g. specifying twice a non-default rand.family
for a random effect shared by two submodels).
A (single) list of class HLfit
, as returned by other fitting functions in spaMM. The main difference is that it contains a families
element describing the response families, instead of the family
elements of fitted objects for univariate response.
See further examples in mv
(modelling correlated random effects over the different submodels),
and residVar
.
### Data preparation data(clinics) climv <- clinics (fitClinics <- HLfit(cbind(npos,nneg)~treatment+(1|clinic), family=binomial(),data=clinics)) set.seed(123) climv$np2 <- simulate(fitClinics, type="residual") # ### fits #### Shared random effect (mvfit <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(1|clinic), family=poisson(), fixed=list(lambda=c("1"=1)))), data=climv)) # Two univariate-response independent fits because random effect terms are distinct # (note how two lambda values are set; same syntax for 'init' values): (mvfitind <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(+1|clinic),family=poisson())), data=climv, fixed=list(lambda=c('1'=1,'2'=0.5)))) # '1': (1|clinic); '2': (+1|clinic) #### Specifying fixed (but not init) values in submodels is also possible (maybe not a good idea) # (mvfitfix <- fitmv( # submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic), # family=binomial(),fixed=list(lambda=c('1'=1))), # '1': (1|clinic) # mod2=list(formula=np2~treatment+(+1|clinic),family=poisson(), # fixed=list(lambda=c('1'=0.5)))), # '2': (+1|clinic) # data=climv)) #### Shared fixed effect # Suppose we want to fit the same intercept for the two submodels # (there may be cases where this is meaningful, even if not here). # The original fit has four coefficients corresponding to four columns # of fixed-effect design matrix: head(design_X <- model.matrix(mvfit)) # (Intercept)_1 treatment_1 (Intercept)_2 treatment_2 # [1,] 1 1 0 0 # ... # The three coefficients of the intended model are (say) # "(Intercept)" "treatment_1" "treatment_2" # We build a matrix that relates the original 4 coefficients to these 3 ones: X_4to3 <- matrix(c(1,0,0, 0,1,0, 1,0,0, 0,0,1), nrow=4, ncol=3, byrow=TRUE, dimnames=list(NULL, c("(Intercept)","treatment_1","treatment_2"))) # defined such that design_X %*% X_4to3 will be the design matrix for the intended model, # and the single "(Intercept)" coefficient of the three-parameter model will operate as # a shared estimate of the "(Intercept)_1" and "(Intercept)_2" coefficients # of the original 4-coefficients model, as intended. # To define such matrices, it is *strongly advised* to fit the unconstrained model first, # and to examine the structure of its model matrix (as shown above). # The new fit is obtained by providing the matrix as the 'X2X' argument: (mvfit3 <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(1|clinic), family=poisson(), fixed=list(lambda=c("1"=1)))), X2X = X_4to3, data=climv)) # => the column names of 'X_4to3' are the fixed-effect names in all output. #### Prediction with a residual-dispersion model set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE), x_het=runif(100), y2=runif(100)) (mvfit <- fitmv(list(list(y ~1+(1|grp), family=beta_resp(), resid.model = ~x_het), list(y2 ~1+(1|grp), family=beta_resp())), data= beta_dat)) misspred <- beta_dat[1:3,] misspred$x_het[1] <- NA # missing info for residual variance of first submodel ## => prediction missing when this info is needed: # length(predict(mvfit, newdata=misspred)) # 6 values: missing info not needed for point predictions length(get_residVar(mvfit, newdata=misspred)) # 5 values length(get_respVar(mvfit, newdata=misspred)) # 5 values # Missing info not needed for predVar (**as opposed to respVar**) length(get_predVar(mvfit, newdata=misspred)) # 6 values # # Same logic for interval computations: # dim(attr(predict(mvfit, newdata=misspred, intervals="respVar"),"intervals")) # 5,2 dim(attr(predict(mvfit, newdata=misspred, intervals="predVar"),"intervals")) # 6,2 # # Same logic for simulate(): # length(simulate(mvfit, newdata=misspred)) # 5 as simulation requires residVar
### Data preparation data(clinics) climv <- clinics (fitClinics <- HLfit(cbind(npos,nneg)~treatment+(1|clinic), family=binomial(),data=clinics)) set.seed(123) climv$np2 <- simulate(fitClinics, type="residual") # ### fits #### Shared random effect (mvfit <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(1|clinic), family=poisson(), fixed=list(lambda=c("1"=1)))), data=climv)) # Two univariate-response independent fits because random effect terms are distinct # (note how two lambda values are set; same syntax for 'init' values): (mvfitind <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(+1|clinic),family=poisson())), data=climv, fixed=list(lambda=c('1'=1,'2'=0.5)))) # '1': (1|clinic); '2': (+1|clinic) #### Specifying fixed (but not init) values in submodels is also possible (maybe not a good idea) # (mvfitfix <- fitmv( # submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic), # family=binomial(),fixed=list(lambda=c('1'=1))), # '1': (1|clinic) # mod2=list(formula=np2~treatment+(+1|clinic),family=poisson(), # fixed=list(lambda=c('1'=0.5)))), # '2': (+1|clinic) # data=climv)) #### Shared fixed effect # Suppose we want to fit the same intercept for the two submodels # (there may be cases where this is meaningful, even if not here). # The original fit has four coefficients corresponding to four columns # of fixed-effect design matrix: head(design_X <- model.matrix(mvfit)) # (Intercept)_1 treatment_1 (Intercept)_2 treatment_2 # [1,] 1 1 0 0 # ... # The three coefficients of the intended model are (say) # "(Intercept)" "treatment_1" "treatment_2" # We build a matrix that relates the original 4 coefficients to these 3 ones: X_4to3 <- matrix(c(1,0,0, 0,1,0, 1,0,0, 0,0,1), nrow=4, ncol=3, byrow=TRUE, dimnames=list(NULL, c("(Intercept)","treatment_1","treatment_2"))) # defined such that design_X %*% X_4to3 will be the design matrix for the intended model, # and the single "(Intercept)" coefficient of the three-parameter model will operate as # a shared estimate of the "(Intercept)_1" and "(Intercept)_2" coefficients # of the original 4-coefficients model, as intended. # To define such matrices, it is *strongly advised* to fit the unconstrained model first, # and to examine the structure of its model matrix (as shown above). # The new fit is obtained by providing the matrix as the 'X2X' argument: (mvfit3 <- fitmv( submodels=list(mod1=list(formula=cbind(npos,nneg)~treatment+(1|clinic),family=binomial()), mod2=list(formula=np2~treatment+(1|clinic), family=poisson(), fixed=list(lambda=c("1"=1)))), X2X = X_4to3, data=climv)) # => the column names of 'X_4to3' are the fixed-effect names in all output. #### Prediction with a residual-dispersion model set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE), x_het=runif(100), y2=runif(100)) (mvfit <- fitmv(list(list(y ~1+(1|grp), family=beta_resp(), resid.model = ~x_het), list(y2 ~1+(1|grp), family=beta_resp())), data= beta_dat)) misspred <- beta_dat[1:3,] misspred$x_het[1] <- NA # missing info for residual variance of first submodel ## => prediction missing when this info is needed: # length(predict(mvfit, newdata=misspred)) # 6 values: missing info not needed for point predictions length(get_residVar(mvfit, newdata=misspred)) # 5 values length(get_respVar(mvfit, newdata=misspred)) # 5 values # Missing info not needed for predVar (**as opposed to respVar**) length(get_predVar(mvfit, newdata=misspred)) # 6 values # # Same logic for interval computations: # dim(attr(predict(mvfit, newdata=misspred, intervals="respVar"),"intervals")) # 5,2 dim(attr(predict(mvfit, newdata=misspred, intervals="predVar"),"intervals")) # 6,2 # # Same logic for simulate(): # length(simulate(mvfit, newdata=misspred)) # 5 as simulation requires residVar
This explains how to handle a warning occurring in computation of prediction variance, where the user is directed here.
For Matern or Cauchy correlation models with vanishing scale factor for distances, a warning may be produced when predict.HLfit
(or get_predVar
, etc.) is called with non-NULL newdata
, because a nearly-singular correlation matrix of the random effect is met. To decide what to do in that case, users should compare the values of get_predVar(.)
and get_predVar(., newdata=myfit$data)
(see Example below). In the absence of numerical inaccuracies, The two values should be identical, and in the presence of such inaccuracies, the more reliable value is the first one. In really poor cases, the second syntax may yield negative prediction variances. If users deem the inaccuracies too large, they should use control=list(fix_predVar=TRUE)
in the next call to predict.HLfit
(or get_predVar
, etc.) as shown in the Example. The drawback of this control is that the computation may be slower, and might even exceed memory capacity for large problems (some matrix operations being performed with exact rational arithmetic, which is memory-consuming for large matrices). it is also still experimental, in the sense that I fear that bugs (stop
) may occur. If the user instead chooses control=list(fix_predVar=FALSE)
, the default standard floating-point arithmetic is used, but no warning is issued.
For fix_predVar
left NULL (the default), standard floating-point arithmetic is also used. But in addition (with exceptions: see Details), the warning keeps being issued, and the (possibly costly) computation of the inverse of the correlation matrix is not stored in the fitted model object, hence is repeated for each new prediction variance computation. This is useful to remind users that something needs to be done, but for programming purposes where repeated warnings may be a nuisance, one can use control=list(fix_predVar=NA)
which will issue a warning then perform as control=list(fix_predVar=FALSE)
, i.e. store an approximate inverse so the warning is not issued again. Finally, control=list(fix_predVar=NaN)
will remove the inverse of the correlation matrix from the fitted model object, and start afresh as if the control was NULL.
Nearly-singular correlation matrices of random effects occur in several contexts. For random-slope models, it commonly occurs that the fitted correlation between the random effects for Intercept and slope is 1 or -1, in which case the correlation matrix between these random effects is singular. This led to quite inaccurate computations of prediction variances in spaMM prior to version 3.1.0, but this problem has been fixed.
control=list(fix_predVar=NaN)
may be more appropriate than control=list(fix_predVar=NULL)
when predict.HLfit
is called through code that one cannot control. For this reason, spaMM provides another mode of control of the default. It will convert control=list(fix_predVar=NULL)
to other values when the call stack has call names matching the patterns given byspaMM.getOption("fix_predVar")
(as understood by grep
). Thus if spaMM.getOption("fix_predVar")$"NA"=="MSL|bboptim"
, the default behaviour is that defined by control=list(fix_predVar=NA)
when predict.HLfit
is called through Infusion::MSL
or blackbox::bboptim
. FALSE or TRUE are handled in a similar way.
data("blackcap") fitobject <- corrHLfit(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, ranFix=list(nu=10,rho=0.001)) ## numerically singular C get_predVar(fitobject,newdata=blackcap[6,]) ## => warning => let us apply the recommended procedure: get_predVar(fitobject) get_predVar(fitobject,newdata=fitobject$data) # Negative values again in the second case => easy decision: get_predVar(fitobject,newdata=blackcap[1:6,], control=list(fix_predVar=TRUE)) # now it's accurate # and the accuracy control is stored in the object: get_predVar(fitobject,newdata=blackcap[1:6,]) # Clean and start afresh: get_predVar(fitobject,newdata=blackcap[1:6,], control=list(fix_predVar=NaN))
data("blackcap") fitobject <- corrHLfit(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, ranFix=list(nu=10,rho=0.001)) ## numerically singular C get_predVar(fitobject,newdata=blackcap[6,]) ## => warning => let us apply the recommended procedure: get_predVar(fitobject) get_predVar(fitobject,newdata=fitobject$data) # Negative values again in the second case => easy decision: get_predVar(fitobject,newdata=blackcap[1:6,], control=list(fix_predVar=TRUE)) # now it's accurate # and the accuracy control is stored in the object: get_predVar(fitobject,newdata=blackcap[1:6,]) # Clean and start afresh: get_predVar(fitobject,newdata=blackcap[1:6,], control=list(fix_predVar=NaN))
The fitting functions allow all parameters to be fixed rather than estimated:
* Fixed-effect coefficients can be set by way of the etaFix
argument (linear predictor coefficients) for all fitting functions.
* Random-effect parameters and the phi
parameter of the gaussian and Gamma response families can be set for all fitting function by the fixed
argument, or for some fitting functions by an alternative argument with the same effect (see Details for this confusing feature, but using fixed
uniformly is simpler).
* The ad-hoc dispersion parameter of some response families (COMPoisson
, negbin1
, negbin2
, beta_resp
, betabin
and possibly future ones) can be fixed using the ad-hoc argument of such families rather than by fixed
.
etaFix is a list with single documented element beta
, which should be a vector of (a subset of) the coefficients () of the fixed effects, with names as shown in a fit without such given values. If REML is used to fit random effect parameters, then
etaFix
affects by default the REML correction for estimation of dispersion parameters, which depends only on which coefficients are estimated rather than given. This default behaviour will be overridden whenever a non-null
REMLformula
is provided to the fitting functions (see Example). Alternatively, with a non-NULL etaFix$beta
, REML can also be performed as if all coefficients were estimated, by adding attribute
keepInREML=TRUE
to etaFix$beta
; in that case the REML computation is by default that implied by the fixed effects in the full model formula, unless a non-default REMLformula
is also used.
The older equivalent for the fixed
argument is ranFix
for HLfit
and corrHLfit
, and ranPars
for HLCor
. Do not use both one such argument and fixed
in a call. This older diversity of names was confusing, but its logic was that ranFix
allows one to fix parameters that HLfit
and corrHLfit
would otherwise estimate, while ranPars
can be used to set correlation parameters that HLCor
does not estimate but nevertheless requires (e.g., Matérn parameters).
Theses arguments for fixing random-effect parameters all have a common syntax. They is a list, with the following possible elements, whose nature is further detailed below:
* phi (variance of residual error, for gaussian and Gamma HGLMs),
* lambda (random-effect variances, except for random-coefficient terms),
* ranCoefs (random-coefficient parameters),
* corrPars (correlation parameters, when handled by the fitting function).
* Individual correlation parameters such as rho, nu, Nugget, ARphi... are also possible top-level elements of the list when there is no ambiguity as to which random effect these correlation parameters apply. This syntax was conceived when spaMM
handled a single spatial random effect, and it is still convenient when applicable, but it should not be mixed with corrPars
element usage.
phi may be a single value or a vector of the same length as the response vector (the number of rows in the data
, once non-informative rows are removed).
lambda may be a single value (if there is a single random effect, or a vector allowing to specify unambiguously variance values for some random effect(s). It can thus take the form lambda=c(NA,1)
or lambda=c("2"=1)
(note the name) to assign a value only to the variance of the second of two random effects.
ranCoefs is a list
of numeric vectors, each numeric vector specifying the variance and correlation parameters for a random-coefficient term. As for lambda
, it may be incomplete, using names to specify the random effect to which the parameters apply. For example, to assign variances values 3 and 7, and correlation value -0.05, to the second random effect in a model formula, one can use ranCoefs=list("2"=c(3,-0.05,7))
(note the name). The elements of each vector are variances and correlations, matching those of the printed summary of a fit. The order of these elements must be the order of the lower.tri
of a covariance matrix, as shown e.g. bym2 <- matrix(NA, ncol=2,nrow=2); m2[lower.tri(m2,diag=TRUE)] <- seq(3); m2
.fitme
accepts partially fixed parameters for a random coefficient term, e.g.,ranCoefs=list("2"=c(NA,-0.05,NA))
, although this may not mix well with some obscure options, such ascontrol=list(refit=list(ranCoefs=TRUE))
which will ignore the fixed values.
GxE
shows how to use partially-fixed ranCoefs
to fit different variances for different levels of a factor.
corrPars is a list, and it may also be incomplete, using names to specify the affected random effect as shown for lambda
and ranCoefs
. For example, ranFix=list(corrPars=list("1"=list(nu=0.5)))
makes explicit that nu=0.5
applies to the first ("1"
) random effect in the model formula. Its elements may be the correlation parameters of the given random effect. For the Matérn model, these are the correlation parameters rho
(scale parameter(s)), nu
(smoothness parameter), and (optionally) Nugget
(see Matern
). The rho
parameter can itself be a vector with different values for different geographic coordinates.
For the adjacency
model, the only correlation parameter is a scalar rho
(see adjacency
).
For the AR1
model, the only correlation parameter is a scalar ARphi
(see AR1
).
Consult the documentation for other types of random effects, such as Cauchy
or IMRF
, for any information missing here.
## Not run: data("wafers") # Fixing random-coefficient parameters: fitme(y~X1+(X2|batch), data=wafers, fixed=list(ranCoefs=list("1"=c(2760, -0.1, 1844)))) ## HLfit syntax for the same effect (except that REML is used here) # HLfit(y~X1+(X2|batch), data=wafers, ranFix=list(ranCoefs=list("1"=c(2760, -0.1, 1844)))) ### Fixing coefficients of the linear predictor: # ## ML fit # fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), etaFix=list(beta=c("(Intercept)"=5.61208))) # ## REML fit # Evaluation of restricted likelihood depends on which fixed effects are estimated, # so simply fixing the coefficients to their REML estimates will not yield # the same REML fits, as see by comparing the next two fits: # unconstr <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML") # # Second fit is different from 'unconstr' despite the same fixed-effects: naive <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML", etaFix=list(beta=fixef(unconstr))) # # Using REMLformula to obtain the same REML fit as the unconstrained one: fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML", etaFix=list(beta=fixef(unconstr)), REMLformula=y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch)) data("Loaloa") # Fixing some Matern correlation parameters, in fitme(): fitme(cbind(npos,ntot-npos) ~ elev1 +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),fixed=list(nu=0.5,Nugget=2/7)) # Fixing all mandatory Matern correlation parameters, in HLCor(): HLCor(cbind(npos,ntot-npos) ~ elev1 + Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranPars=list(nu=0.5,rho=0.7)) ## End(Not run)
## Not run: data("wafers") # Fixing random-coefficient parameters: fitme(y~X1+(X2|batch), data=wafers, fixed=list(ranCoefs=list("1"=c(2760, -0.1, 1844)))) ## HLfit syntax for the same effect (except that REML is used here) # HLfit(y~X1+(X2|batch), data=wafers, ranFix=list(ranCoefs=list("1"=c(2760, -0.1, 1844)))) ### Fixing coefficients of the linear predictor: # ## ML fit # fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), etaFix=list(beta=c("(Intercept)"=5.61208))) # ## REML fit # Evaluation of restricted likelihood depends on which fixed effects are estimated, # so simply fixing the coefficients to their REML estimates will not yield # the same REML fits, as see by comparing the next two fits: # unconstr <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML") # # Second fit is different from 'unconstr' despite the same fixed-effects: naive <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML", etaFix=list(beta=fixef(unconstr))) # # Using REMLformula to obtain the same REML fit as the unconstrained one: fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log), method="REML", etaFix=list(beta=fixef(unconstr)), REMLformula=y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch)) data("Loaloa") # Fixing some Matern correlation parameters, in fitme(): fitme(cbind(npos,ntot-npos) ~ elev1 +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),fixed=list(nu=0.5,Nugget=2/7)) # Fixing all mandatory Matern correlation parameters, in HLCor(): HLCor(cbind(npos,ntot-npos) ~ elev1 + Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranPars=list(nu=0.5,rho=0.7)) ## End(Not run)
fixedLRT
performs a likelihood ratio (LR) test between two models, the “full” and the “null” models,
currently differing only in their fixed effects. Parametric bootstrap p-values can be computed, either using the raw bootstrap distribution of the likelihood ratio, or a bootstrap estimate of the Bartlett correction of the LR statistic.
This function differs from LRT
in its arguments (model fits for LRT
, versus all arguments required to fit the models for fixedLRT
), and in the format of its return value.
fixedLRT(null.formula, formula, data, method, HLmethod = method, REMLformula = NULL, boot.repl=0, control="DEPRECATED", control.boot="DEPRECATED", fittingFunction, seed=NULL, resp_testfn = NULL, weights.form = NULL, ...)
fixedLRT(null.formula, formula, data, method, HLmethod = method, REMLformula = NULL, boot.repl=0, control="DEPRECATED", control.boot="DEPRECATED", fittingFunction, seed=NULL, resp_testfn = NULL, weights.form = NULL, ...)
null.formula |
Either a |
formula |
Either a |
data |
A data frame containing the variables in the model. |
method |
A method to fit the full and null models.
See Also feasible, but more tricky, and not really recommended (see Rousset and Ferdy, 2014), is |
HLmethod |
Kept for back-compatibility. Same as |
REMLformula |
a formula specifying the fixed effects which design matrix is used in the REML correction for the estimation of dispersion parameters, if these are estimated by REML. This formula is by default that for the *full* model. |
weights.form |
Specification of prior weights by a one-sided formula: use |
boot.repl |
the number of bootstrap replicates. |
control |
Deprecated. |
control.boot |
Deprecated. |
fittingFunction |
Character string giving the function used to fit each model: either |
seed |
Passed to |
resp_testfn |
See argument |
... |
Further arguments passed to or from other methods; presently, additional arguments passed to fitting functions. |
Comparison of REML fits is a priori not suitable for performing likelihood ratio tests. Nevertheless, it is possible to contrive them for testing purposes (Welham & Thompson 1997). This function generalizes some of Wehlam & Thompson's methods to GLMMs.
See Details in LRT
for details of the bootstrap procedures.
An object of class fixedLRT
, actually a list with as-yet unstable format, but here with typical elements (depending on the options)
fullfit |
the HLfit object for the full model; |
nullfit |
the HLfit object for the null model; |
LRTori |
A likelihood ratio chi-square statistic |
LRTprof |
Another likelihood ratio chi-square statistic, after a profiling step, if any. |
df |
the number of degrees of freedom of the test. |
trace.info |
Information on various steps of the computation. |
and, if a bootstrap was performed, the additional elements described in LRT
.
Rousset F., Ferdy, J.-B. (2014) Testing environmental and genetic effects in the presence of spatial autocorrelation. Ecography, 37: 781-790. doi:10.1111/ecog.00566
Welham, S. J., and Thompson, R. (1997) Likelihood ratio tests for fixed model terms using residual maximum likelihood, J. R. Stat. Soc. B 59, 701-714.
See LRT
for simular tests with a different interface, and perhaps as_LMLT
for access to a different testing approach for LMMs, implemented in lmerTest::contest
.
if (spaMM.getOption("example_maxtime")>1.9) { data("blackcap") ## result comparable to the corrHLfit examples based on blackcap fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap) } if (spaMM.getOption("example_maxtime")>156) { ## longer version with bootstrap fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap, boot.repl=100, seed=123) }
if (spaMM.getOption("example_maxtime")>1.9) { data("blackcap") ## result comparable to the corrHLfit examples based on blackcap fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap) } if (spaMM.getOption("example_maxtime")>156) { ## longer version with bootstrap fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap, boot.repl=100, seed=123) }
A set of data on airfreight breakage. Data are given on 10 air shipments, each carrying 1000 ampules of some substance. For each shipment, the number of ampules found broken upon arrival, and the number of times the shipments were transferred from one aircraft to another, are recorded.
data("freight")
data("freight")
The data frame includes 10 observations on the following variables:
number of ampules found broken upon arrival.
number of times the shipments were transferred from one aircraft to another.
Shipment identifier.
The data set is reported by Kutner et al. (2003) and used by Sellers & Shmueli (2010) to illustrate COMPoisson analyses.
Kutner MH, Nachtsheim CJ, Neter J, Li W (2005, p. 35). Applied Linear Regression Models, Fourth Edition. McGraw-Hill.
Sellers KF, Shmueli G (2010) A Flexible Regression Model for Count Data. Ann. Appl. Stat. 4: 943–961
## see ?COMPoisson for examples
## see ?COMPoisson for examples
This function is similar to get_predVar
except that is uses a bootstrap procedure to correct for bias in the evaluation of the prediction variance.
get_cPredVar(pred_object, newdata, nsim, seed, type = "residual", variances=NULL, nb_cores = NULL, fit_env = NULL, sim_object=pred_object)
get_cPredVar(pred_object, newdata, nsim, seed, type = "residual", variances=NULL, nb_cores = NULL, fit_env = NULL, sim_object=pred_object)
pred_object |
an object of class |
newdata |
passed to |
nsim |
passed to |
seed |
passed to |
type |
passed to |
variances |
NULL or list; |
nb_cores |
integer: number of cores to use for parallel computation of bootstrap. The default is |
fit_env |
For parallel computations: an environment containing objects to be passed to the cores. They should have the same name in |
sim_object |
an object of class |
The result provided by get_cPredVar
is similar to the CMSEP (Conditional Mean Standard Error of Prediction) introduced by Booth and Hobert (1998; “B&H”). This paper is known for pointing the importance of using conditional variances when they differ from unconditional ones. This is hard to miss in spatial models, where the relevant prediction variance typically depends on the variance of random effects conditional on the data. Thus, the alternative function get_predVar
already accounts for this and returns a prediction variance that depends on a joint covariance of fixed-effect estimates and of random effects given the data.
B&H also used a conditional bootstrap procedure to correct for some bias. get_cPredVar
implements a similar procedure, in contrast to get_predVar
. Their conditional bootstrap procedure is not applicable for autocorrelated random effects, and parametric bootstrapping of the residuals of the fitted model (as implied by the default value of argument type
) is used instead here. Apart from this difference, the returned value includes exactly the same terms as those discussed by B&H: their “naive estimate” and its bootstrap correction
, their correction
for uncertainty in fixed-effect coefficients, and their correction
for uncertainty in dispersion parameters.
This use of the bootstrap does not account for uncertainty in correlation parameters “outer-optimized” by fitme
or corrHLfit
, because the correlation parameters are fixed when the model is refitted on the bootstrap replicates. Even if it the correlation parameters were refitted, the full computation would not be sufficient to account for uncertainty in them. To account for uncertainty in correlation parameters, one should rather perform a parametric bootstrap of the full model (typically using spaMM_boot(., type="residual")
), which may take much more time.
The “naive estimate” is not generally an estimate of anything uniquely defined by the model parameters: for correlated random effects, it depends on the “root” of the correlation matrix of the random effects, which is not unique. Thus
is not unique, and may differ for example for equivalent fits by sparse-precision methods vs. other methods. Nevertheless,
attr(cpredvar,"info")$naive
does recover published values in the Examples below, as they involve no correlation matrix.
A vector of prediction variances, with an attribute info
which is an environment containing variables:
SEs |
the standard errors of the estimates (which are those of the bootstrap replicates) |
bias |
the bias term |
maive |
B&H's “naive” |
Booth, J.G., Hobert, J.P. (1998) Standard errors of prediction in generalized linear mixed models. J. Am. Stat. Assoc. 93: 262-272.
## Not run: if(requireNamespace("rsae", quietly = TRUE)) { # LMM example from Booth & Hobert 1998 JASA data("landsat", package = "rsae") fitCorn <- fitme(HACorn ~ PixelsCorn + PixelsSoybeans + (1|CountyName),data=landsat[-33,]) newXandZ <- unique(data.frame(PixelsCorn=landsat$MeanPixelsCorn, PixelsSoybeans=landsat$MeanPixelsSoybeans, CountyName=landsat$CountyName)) (cpredvar <- get_cPredVar(fitCorn, newdata=newXandZ, nsim=200L, seed=123)) # serial computation (cpredvar <- get_cPredVar(fitCorn, newdata=newXandZ, nsim=200L, seed=123, nb_cores=parallel::detectCores(logical=FALSE)-1L, fit_env=list2env(list(newXandZ=newXandZ)))) } # GLMM example from Booth & Hobert 1998 JASA data(clinics) fitClinics <- HLfit(cbind(npos,nneg)~treatment+(1|clinic),family=binomial(),data=clinics) # (get_cPredVar(fitClinics, newdata=clinics[1:8,], nsim=200L, seed=123)) # serial computation (get_cPredVar(fitClinics, newdata=clinics[1:8,], nsim=200L, seed=123, nb_cores=parallel::detectCores(logical=FALSE)-1L, fit_env=list2env(list(clinics=clinics)))) ## End(Not run)
## Not run: if(requireNamespace("rsae", quietly = TRUE)) { # LMM example from Booth & Hobert 1998 JASA data("landsat", package = "rsae") fitCorn <- fitme(HACorn ~ PixelsCorn + PixelsSoybeans + (1|CountyName),data=landsat[-33,]) newXandZ <- unique(data.frame(PixelsCorn=landsat$MeanPixelsCorn, PixelsSoybeans=landsat$MeanPixelsSoybeans, CountyName=landsat$CountyName)) (cpredvar <- get_cPredVar(fitCorn, newdata=newXandZ, nsim=200L, seed=123)) # serial computation (cpredvar <- get_cPredVar(fitCorn, newdata=newXandZ, nsim=200L, seed=123, nb_cores=parallel::detectCores(logical=FALSE)-1L, fit_env=list2env(list(newXandZ=newXandZ)))) } # GLMM example from Booth & Hobert 1998 JASA data(clinics) fitClinics <- HLfit(cbind(npos,nneg)~treatment+(1|clinic),family=binomial(),data=clinics) # (get_cPredVar(fitClinics, newdata=clinics[1:8,], nsim=200L, seed=123)) # serial computation (get_cPredVar(fitClinics, newdata=clinics[1:8,], nsim=200L, seed=123, nb_cores=parallel::detectCores(logical=FALSE)-1L, fit_env=list2env(list(clinics=clinics)))) ## End(Not run)
get_inits_from_fit
is an extractor of some fitted values from a fit in a convenient format to initiate a next fit.
get_inits_from_fit(from, template = NULL, to_fn = NULL, inner_lambdas=FALSE)
get_inits_from_fit(from, template = NULL, to_fn = NULL, inner_lambdas=FALSE)
from |
Fit object (inheriting from class |
template |
Another fit object. Usage with a |
to_fn |
NULL or character: the name of the function to be used the next fit. If NULL, taken from |
inner_lambdas |
Boolean;
Whether the output should include estimates of the dispersion parameters estimated by the iterative methods implemented in |
A list with elements
init , init.corrHLfit
|
(depending on the fitting function) giving initial values for outer-optimization; |
init.HLfit |
giving initial values for the iterative algorithms in
|
get_ranPars
and VarCorr
.
## Not run: data("blackcap") (corrhlfit <- corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML")) inits <- get_inits_from_fit(corrhlfit, to_fn = "fitme") (fitfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=inits[["init"]])) inits <- get_inits_from_fit(corrhlfit, template = fitfit) fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=inits[["init"]]) # In these examples, inits$init.HLfit is useless # as it is ignored when LMMs are fitted by fitme(). ## End(Not run)
## Not run: data("blackcap") (corrhlfit <- corrHLfit(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML")) inits <- get_inits_from_fit(corrhlfit, to_fn = "fitme") (fitfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=inits[["init"]])) inits <- get_inits_from_fit(corrhlfit, template = fitfit) fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, init=inits[["init"]]) # In these examples, inits$init.HLfit is useless # as it is ignored when LMMs are fitted by fitme(). ## End(Not run)
get_matrix
is a first attempt at a unified extractor of various matrices from a fit. All augmented matrices follow (Henderson's) block order (upper blocks: X,Z; lower blocks: 0,I).
get_ZALMatrix
returns the design matrix for the random effects .
get_matrix(object, which="model.matrix", augmented=TRUE, ...) get_ZALMatrix(object, force_bind=TRUE)
get_matrix(object, which="model.matrix", augmented=TRUE, ...) get_ZALMatrix(object, force_bind=TRUE)
object |
An object of class |
augmented |
Boolean; whether to return a matrix for all model coefficients (augmented matrix for fixed-effects coefficients and random-effect predictions) or a matrix only for fixed effects. Not operative for all |
which |
Which element to extract. For |
force_bind |
Boolean; with the non-default value |
... |
Other arguments that may be needed in some future versions of |
(Given the pain that it is to write maths in R documentation files, readers are gently asked to be tolerant about any imperfections of the following).
Model coefficients estimates of a (weighted) linear model can be written as (X'WX)X'Wy where X is the design matrix for fixed effects, W a diagonal weight matrix, and y the response vector. In a linear mixed model, the same expression holds in terms of Henderson's augmented design matrix, of an augmented (still diagonal) weight matrix, and of an augmented response vector. For GLMMs and hierarchical GLMs generally, the solution of each step of the iteratively reweighted least squares algorithm again has the same expression in terms of appropriately defined augmented matrices and vectors.
get_matrix
returns, for given values of the which
argument, the following matrices from the model fit:"AugX"
: X;"wei_AugX"
: WX;"wAugX"
: (W)X;
"left_ginv"
: (X'WX)X'W (the name stems from the fact that it is generalized inverse, denoted X
, since XX
X=X, and it is a left one, since X
X is an identity matrix when X has full rank);
"hat_matrix"
: XX=X (X'WX)
X'W;
"fixef_left_ginv"
: same as "left_ginv"
but for the fixed-effect design matrix only (not to be confused with the corresponding block of "left_ginv"
);"beta_v_cov"
: joint covariance matrix of estimates/predictions of fixed-effect coefficients and random effects (the v in random-effects
).
"v_condcov"
: covariance matrix of predictions of random effects (the v in random-effects
) given fixed-effect coefficients.
A matrix, possibly in sparseMatrix
format.
vcov
for the variance-covariance matrix of the fixed-effects coefficients, and Corr
for correlation matrices of random effects.
get_fittedPars
returns estimated parameters.
get_ranPars
returns various subsets of random-effect parameters (correlation or variance parameters), as controlled by its which
argument.
It is one of several extractors for fixed or estimated parameters of different classes of parameters, for which a quick guide is
get_ranPars
: for random-effect parameters, excluding residual dispersion (with a subtlety for corrFamily
models: see Details);VarCorr
: alternative extractor for random-effect (co)variance and optionally residual variance, in a data frame format;residVar
: for residual variance parameters, family dispersion parameters, or information about residual variance models;get_residVar
: alternative extractor of residual variances with different features inherited from get_predVar
;get_inits_from_fit
: extracts estimated parameters from a fit, in a different format from get_fittedPars
.
remove_from_parlist
removes elements from a list of parameters, and from its type
attribute.
get_fittedPars(object, partial_rC="rm", phiPars=TRUE) get_ranPars(object, which=NULL, verbose=TRUE, lambda_names = "Group.Term", ...) remove_from_parlist(parlist, removand=NULL, rm_names=names(unlist(removand)))
get_fittedPars(object, partial_rC="rm", phiPars=TRUE) get_ranPars(object, which=NULL, verbose=TRUE, lambda_names = "Group.Term", ...) remove_from_parlist(parlist, removand=NULL, rm_names=names(unlist(removand)))
object |
An object of class |
partial_rC |
Controls handling of partially-fixed random coefficients. The default as set by |
phiPars |
Boolean: whether to include the parameters of any residual-dispersion model for |
which |
NULL or character string. Use |
... |
Other arguments that may be needed by some method. |
verbose |
Boolean: Whether to print some notes. |
parlist |
A list of parameters. see Details. |
removand |
Optional. A list of parameters to be removed from |
rm_names |
Names of parameters to be removed from |
lambda_names |
By default the names of the |
For heteroscedastic random effects, such as conditional autoregressive models, the variance parameter “lambda” refers to a common scaling coefficient. For other random-effect models, “lambda” typically refers to the single variance parameter.
remove_from_parlist
is designed to manipulate structured lists of parameters, such as a list with elements phi
, lambda
, and corrPars
, the latter being itself a list structured as the return value of get_ranPars(.,which="corrPars")
. parlist
may have an attribute type
, also with elements phi
, lambda
, and corrPars
... If given, removand
must have the same structure (but typically not all the elements of parlist
); otherwise, rm_names
must have elements which match names of unlist(names(parlist))
.
If a corrFamily parameter is fixed through the formula term, as in ARp(1|time, p=3, fixed=c(p2=0))
, the fixed parameter is not considered a model parameter and get_ranPars
will not extract it from the object. However, the parameter will be extracted if it has been fixed through fitme
's fixed
argument rather than through the formula term (see example in ARp
).
get_fittedPars
returns a list of model parameters, with possible elements: beta
(fixed-effect coefficients); lambda
, phi
, ranCoefs
and corrPars
(same meaning as in fixed
parameters); hyper
, for multIMRF
models;
the residual-dispersion parameters beta_prec
, NB_shape
and COMP_nu
when they are single scalars;
and rdisPars
for more complex residual-dispersion parameters. See the specific resid.model
and phi-resid.model
documentations for the rdisPars
format, dependent on the nature of the residual-dispersion parameter being modelized. Use residVar(., which="fam_parm")
to extract the vector of fitted values of the dispersion parameter.
get_ranPars(.,which="corrPars")
returns a (possibly nested) list of correlation parameters (or NULL if there is no such parameter). Top-level elements correspond to the different random effects. The list has a "type"
attribute having the same nested-list structure and describing whether and how the parameters where fitted: "fix"
means they where fixed, not fitted; "var"
means they were fitted by HLfit
's specific algorithms; "outer"
means they were fitted by a generic optimization method.
get_ranPars(.,which="lambda")
returns a vector of variance values, one per random effect, including both fixed, “outer”- and “inner”-optimized ones. The variances of random-coefficients terms with correlation parameters are not included.
get_ranPars(.,which="outer_lambda")
returns only “outer”-optimized variance parameters, ignoring those fitted by HLfit
's specific algorithms.
get_ranPars(.,which=NULL)
(the default) is not fully defined. It returns a list including the results of which="lambda"
and which="corrPars"
, but possibly other elements too.
get_ranPars(.,which="fitted")
is designed to provide fitted parameters with respect to which an information matrix is to be calculated (using numDeriv. It excludes fixed values, and has no type
attribute.
get_ranPars(. which="ranef_var")
(experimental) returns a list with elements
Var
same as get_ranPars(.,which="lambda")
lembda_est
A vector of variance values, one for each level of each random effect
outer
A vector or outer-optimized variance values, as returned by get_ranPars(.,which="outer_lambda")
Other elements, subject to change in later versions.
remove_from_parlist
returns a list of model parameters with given elements removed, and likewise for its (optional) type
attribute. See Details for context of application.
See get_fittedPars
, VarCorr
, residVar
, get_residVar
, or get_inits_from_fit
as described in the quick guide above.
data("wafers") m1 <- HLfit(y ~X1+X2+(1|batch), resid.model = ~ 1, data=wafers, method="ML") get_ranPars(m1,which="corrPars") # NULL since no correlated random effect parlist1 <- list(lambda=1,phi=2,corrPars=list("1"=list(rho=3,nu=4),"2"=list(rho=5))) parlist2 <- list(lambda=NA,corrPars=list("1"=list(rho=NA))) # values of elements do not matter remove_from_parlist(parlist1,parlist2) ## same result as: remove_from_parlist(parlist1,rm_names = names(unlist(parlist2)))
data("wafers") m1 <- HLfit(y ~X1+X2+(1|batch), resid.model = ~ 1, data=wafers, method="ML") get_ranPars(m1,which="corrPars") # NULL since no correlated random effect parlist1 <- list(lambda=1,phi=2,corrPars=list("1"=list(rho=3,nu=4),"2"=list(rho=5))) parlist2 <- list(lambda=NA,corrPars=list("1"=list(rho=NA))) # values of elements do not matter remove_from_parlist(parlist1,parlist2) ## same result as: remove_from_parlist(parlist1,rm_names = names(unlist(parlist2)))
get_RLRsim_args
extracts a list of arguments suitable for a call to RLRsim::RLRTSim()
or RLRsim::LRTSim()
. These functions use an efficient simulation procedure to compute restricted or marginal likelihood ratio tests, respectively, comparing a fixed-effect model and a mixed-effect model with one random effect. They are notably used to test for the presence of one random effect, although the models compared by marginal likelihood (LRTSim()
) may differ both in their random and in their fixed effects (as shown in the Example). The tests are exact for small samples (up to simulation error) for LMMs with no free parameters in the random effect (beyond the variance being tested), so not for time-series or spatial models with estimated correlation parameters. Heteroscedasticity of the residuals or of the random effect variance are also not taken into account by the simulation procedure (see Value field below for an hint why this is so).
get_RLRTSim_args
is the older extractor, originally for RLRsim::RLRTSim()
only, now handling also ML fits with a warning (though the possible absence of the nullfit
argument will result in an error).
get_RLRsim_args(fullfit, nullfit, verbose=TRUE, REML=NA, ...) get_RLRTSim_args(object, verbose=TRUE, ...)
get_RLRsim_args(fullfit, nullfit, verbose=TRUE, REML=NA, ...) get_RLRTSim_args(object, verbose=TRUE, ...)
object , fullfit
|
An object of class |
nullfit |
Same for the less complete model; required only for (marginal) LR test, as opposed to restricted LR test. |
verbose |
NA or boolean; Whether to display some message or not. |
REML |
For programming purposes, not documented. |
... |
Additional arguments (currently not used). |
If the models compared do not differ in their fixed effects, under the null hypothesis there is a probability mass P for a zero likelihood ratio, and the distribution of p-values can be uniform only on the range (0,1-P). If the fixed effects differ (as handled by RLRsim::LRTSim()
), this does not occur.
A list of arguments for a call to RLRsim::RLRTSim()
or RLRsim::LRTSim()
. The main arguments are the design matrix for the fixed effects, and the ZA matrix and L detailed in random-effects
(here represented by the Z
and sqrt.Sigma
elements). The models handled by the testing procedure are the ones that are sufficiently characterized by these two matrices. LRTSim
additionally requires q
, the difference in number of parameters of fixed effects between the models.
The inconsistent capitalisation of 's' in the function names is consistent with the inconsistencies in the RLRsim
package.
Crainiceanu, C. and Ruppert, D. (2004) Likelihood ratio tests in linear mixed models with one variance component, Journal of the Royal Statistical Society: Series B,66,165–185.
The bootstrap procedure in LRT
is more general but slower. It appears to provide results quite similar to those of RLRsim when both are applicable.
## Not run: ## Derived from example in RLRsim::LRTSim set.seed(123) dat <- data.frame(g = rep(1:10, e = 10), x = (x<-rnorm(100)), y = 0.1 * x + rnorm(100)) m <- fitme(y ~ x + (1|g), data=dat) m0 <- fitme(y ~ 1, data=dat) (obs.LRT <- 2*(logLik(m)-logLik(m0))) args <- get_RLRsim_args(m,m0) sim.LRT <- do.call(RLRsim::LRTSim, args ) (RLRpval <- (sum(sim.LRT >= obs.LRT) + 1) / (length(sim.LRT) + 1)) ## comparable test using LRT(): # (bootpval <- LRT(m,m0, boot.repl = 199L)$rawBootLRT$p_value) ## End(Not run)
## Not run: ## Derived from example in RLRsim::LRTSim set.seed(123) dat <- data.frame(g = rep(1:10, e = 10), x = (x<-rnorm(100)), y = 0.1 * x + rnorm(100)) m <- fitme(y ~ x + (1|g), data=dat) m0 <- fitme(y ~ 1, data=dat) (obs.LRT <- 2*(logLik(m)-logLik(m0))) args <- get_RLRsim_args(m,m0) sim.LRT <- do.call(RLRsim::LRTSim, args ) (RLRpval <- (sum(sim.LRT >= obs.LRT) + 1) / (length(sim.LRT) + 1)) ## comparable test using LRT(): # (bootpval <- LRT(m,m0, boot.repl = 199L)$rawBootLRT$p_value) ## End(Not run)
Base fitting functions in R will seek variables in the environment where the formula
was defined (i.e., typically in the global environment), if they are not in the data
. This increases the memory size of fit objects (as the formula and attached environment are part of such objects). This also easily leads to errors (see example in the discussion of update.HLfit
). Indeed Chambers (2008, p.221), after describing how the environment is defined, comments that “Where clear and trustworthy software is a priority, I would personally avoid such tricks. Ideally, all the variables in the model frame should come from an explicit, verifiable data source...”. Fitting functions in spaMM try to adhere to such a principle, as they assume by default that all variables from the formula
should be in the data
argument (and then, one never needs to specify “data$
” in the formula
.. spaMM implements this by default by stripping the formula environment from any variable. It is also possible to assign a given environment to the formula, through the control control.HLfit$formula_env
: see Examples.
The variables defining the prior.weights
should also be in the data
. However, the implementation of the prior.weights
argument has limitations that can be overcome by using the more recently introduced weights.formula
argument of spaMM fitting functions (see Examples, where this is also compared with stats::lm
's handling of its weights
argument).
However, variables used in other arguments such as ranFix
are looked up neither in the data nor in the formula environment, but in the calling environment as usual.
Chambers J.M. (2008) Software for data analysis: Programming with R. Springer-Verlag New York
####### Controlling the formula environment set.seed(123) d2 <- data.frame(y = seq(10)/2+rnorm(5)[gl(5,2)], x1 = sample(10), grp=gl(5,2), seq10=seq(10)) # Using only variables in the data: basic usage # HLfit(y ~ x1 + seq10+(1|grp), data = d2) # is practically equivalent to HLfit(y ~ x1 + seq10+(1|grp), data = d2, control.HLfit = list(formula_env=list2env(list(data=d2)))) # # The 'formula_env' avoids the need for the 'seq10' variable: HLfit(y ~ x1 + I(seq_len(nrow(data)))+(1|grp), data = d2, control.HLfit = list(formula_env=list2env(list(data=d2)))) # # Internal imprementation exploits partial matching of argument names # so that this can also be in 'control' if 'control.HLfit' is absent: fitme(y ~ x1 + I(seq_len(nrow(data)))+(1|grp), data = d2, control = list(formula_env=list2env(list(data=d2)))) ####### Prior-weights misery data("hills", package="MASS") (fit <- lm(time ~ dist + climb, data = hills, weights=1/dist^2)) # same as (fit <- fitme(time ~ dist + climb, data = hills, prior.weights=1/dist^2, method="REML")) # possible calls: (fit <- fitme(time ~ dist + climb, data = hills, prior.weights=quote(1/dist^2))) (fit <- fitme(time ~ dist + climb, data = hills, prior.weights= 1/hills$dist^2)) (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ 1/dist^2)) (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ I(1/dist^2))) # Also syntactically correct since 'dist' is found in the data: (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ rep(2,length(dist)))) #### Programming with prior weights: ## Different ways of passing prior weights to fitme() from another function: wrap_as_form <- function(weights.form) { fitme(time ~ dist + climb, data = hills, weights.form=weights.form) } wrap_as_pw <- function(prior.weights) { fitme(time ~ dist + climb, data = hills, prior.weights=prior.weights) } wrap_as_dots <- function(...) { fitme(time ~ dist + climb, data = hills,...) } ## Similarly for lm: wrap_lm_as_dots <- function(...) { lm(time ~ dist + climb, data = hills, ...) } wrap_lm_as_arg <- function(weights) { lm(time ~ dist + climb, data = hills,weights=weights) } ## Programming errors with stats::lm(): pw <- rep(1e-6,35) # or even NULL (fit <- wrap_lm_as_arg(weights=pw)) # catches weights from global envir! (fit <- lm(time ~ dist + climb, data = hills, weights=pw)) # idem! (fit <- lm(time ~ dist + climb, data = hills, weights=hills$pw)) # fails silently - no $pw in 'hills' (fit <- wrap_lm_as_dots(weights=hills$pw)) # idem! (fit <- wrap_lm_as_arg(weights=hills$pw)) # idem! ## Safer spaMM results: try(fit <- wrap_as_pw(prior.weights= pw)) # correctly catches problem try(fit <- wrap_as_dots(prior.weights=hills$pw)) # correctly catches problem (fit <- wrap_as_dots(prior.weights=1/dist^2)) # correct (fit <- wrap_as_dots(prior.weights=quote(1/dist^2))) # correct ## But 'prior.weights' limitations: try(fit <- wrap_as_pw(prior.weights= 1/hills$dist^2)) # fails (stop) try(fit <- wrap_as_pw(prior.weights= 1/dist^2)) # fails (stop) try(fit <- wrap_as_pw(prior.weights= quote(1/dist^2))) # fails (stop) ## Problems all solved by using 'weights.form': try(fit <- wrap_as_form(weights.form= ~ pw)) # correctly catches problem (fit <- wrap_as_form(weights.form= ~1/dist^2)) # correct (fit <- wrap_as_form(weights.form= ~1/hills$dist^2)) # correct (fit <- wrap_as_dots(weights.form= ~ 1/dist^2)) # correct rm("pw")
####### Controlling the formula environment set.seed(123) d2 <- data.frame(y = seq(10)/2+rnorm(5)[gl(5,2)], x1 = sample(10), grp=gl(5,2), seq10=seq(10)) # Using only variables in the data: basic usage # HLfit(y ~ x1 + seq10+(1|grp), data = d2) # is practically equivalent to HLfit(y ~ x1 + seq10+(1|grp), data = d2, control.HLfit = list(formula_env=list2env(list(data=d2)))) # # The 'formula_env' avoids the need for the 'seq10' variable: HLfit(y ~ x1 + I(seq_len(nrow(data)))+(1|grp), data = d2, control.HLfit = list(formula_env=list2env(list(data=d2)))) # # Internal imprementation exploits partial matching of argument names # so that this can also be in 'control' if 'control.HLfit' is absent: fitme(y ~ x1 + I(seq_len(nrow(data)))+(1|grp), data = d2, control = list(formula_env=list2env(list(data=d2)))) ####### Prior-weights misery data("hills", package="MASS") (fit <- lm(time ~ dist + climb, data = hills, weights=1/dist^2)) # same as (fit <- fitme(time ~ dist + climb, data = hills, prior.weights=1/dist^2, method="REML")) # possible calls: (fit <- fitme(time ~ dist + climb, data = hills, prior.weights=quote(1/dist^2))) (fit <- fitme(time ~ dist + climb, data = hills, prior.weights= 1/hills$dist^2)) (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ 1/dist^2)) (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ I(1/dist^2))) # Also syntactically correct since 'dist' is found in the data: (fit <- fitme(time ~ dist + climb, data = hills, weights.form= ~ rep(2,length(dist)))) #### Programming with prior weights: ## Different ways of passing prior weights to fitme() from another function: wrap_as_form <- function(weights.form) { fitme(time ~ dist + climb, data = hills, weights.form=weights.form) } wrap_as_pw <- function(prior.weights) { fitme(time ~ dist + climb, data = hills, prior.weights=prior.weights) } wrap_as_dots <- function(...) { fitme(time ~ dist + climb, data = hills,...) } ## Similarly for lm: wrap_lm_as_dots <- function(...) { lm(time ~ dist + climb, data = hills, ...) } wrap_lm_as_arg <- function(weights) { lm(time ~ dist + climb, data = hills,weights=weights) } ## Programming errors with stats::lm(): pw <- rep(1e-6,35) # or even NULL (fit <- wrap_lm_as_arg(weights=pw)) # catches weights from global envir! (fit <- lm(time ~ dist + climb, data = hills, weights=pw)) # idem! (fit <- lm(time ~ dist + climb, data = hills, weights=hills$pw)) # fails silently - no $pw in 'hills' (fit <- wrap_lm_as_dots(weights=hills$pw)) # idem! (fit <- wrap_lm_as_arg(weights=hills$pw)) # idem! ## Safer spaMM results: try(fit <- wrap_as_pw(prior.weights= pw)) # correctly catches problem try(fit <- wrap_as_dots(prior.weights=hills$pw)) # correctly catches problem (fit <- wrap_as_dots(prior.weights=1/dist^2)) # correct (fit <- wrap_as_dots(prior.weights=quote(1/dist^2))) # correct ## But 'prior.weights' limitations: try(fit <- wrap_as_pw(prior.weights= 1/hills$dist^2)) # fails (stop) try(fit <- wrap_as_pw(prior.weights= 1/dist^2)) # fails (stop) try(fit <- wrap_as_pw(prior.weights= quote(1/dist^2))) # fails (stop) ## Problems all solved by using 'weights.form': try(fit <- wrap_as_form(weights.form= ~ pw)) # correctly catches problem (fit <- wrap_as_form(weights.form= ~1/dist^2)) # correct (fit <- wrap_as_form(weights.form= ~1/hills$dist^2)) # correct (fit <- wrap_as_dots(weights.form= ~ 1/dist^2)) # correct rm("pw")
Loading these data loads three objects describing a mythical 'Gryphon' population used by Wilson et al. to illustrate mixed-effect modelling in quantitative genetics. These objects are a data frame Gryphon_df
containing the model variables, a genetic relatedness matrix Gryphon_A
, and another data frame Gryphon_pedigree
containing pedigree information (which can be used by some packages to reconstruct the relatedness matrix).
data("Gryphon")
data("Gryphon")
Gryphon_df
is
'data.frame': 1084 obs. of 6 variables: $ ID : int 1029 1299 ...: individual identifier $ sex : Factor w/ 2 levels "1","2": sex, indeed $ year : Factor w/ 34 levels "968","970", ...: birth year $ mother: Factor w/ 429 levels "1","2",..: individual's mother identifier $ BWT : num 10.77 9.3 ...: birth weight $ TARSUS: num 24.8 22.5 12 ...: tarsus length
Gryphon_A
is a genetic relatedness matrix, in sparse matrix format, for 1309 individuals.
Gryphon_pedigree
is
'data.frame': 1309 obs. of 3 variables: $ ID : int 1306 1304 ...: individual identifier $ Dam : int NA NA ...: individual's mother $ Sire: int NA NA ...: individual's father
Wilson AJ, et al. (2010) An ecologist's guide to the animal model. Journal of Animal Ecology 79(1): 13-26. doi:10.1111/j.1365-2656.2009.01639.x
#### Bivariate-response model used as example in Wilson et al. (2010): # joint modelling of birth weight (BWT) and tarsus length (TARSUS). # The relatedness matrix is specified as a 'corrMatrix'. The random # effect 'corrMatrix(0+mv(1,2)|ID)' then represents genetic effects # correlated over traits and individuals (see help("composite-ranef")). # The ...(0+...) syntax avoids contrasts being used in the design # matrix of the random effects, as it would not does make much sense # to represent TARSUS as a contrast to BWT. # The relatedness matrix will be specified through its inverse, # using as_precision(), so that spaMM does not have to find out and # inform the user that using the inverse is better (as is typically # the case for relatedness matrices). But using as_precision() is # not required. See help("algebra") for Details. # The second random effect '(0+mv(1,2)|ID)' represents correlated # environmental effects. Since measurements are not repeated within # individuals, this effect also absorbs all residual variation. The # residual variances 'phi' must then be fixed to some negligible values # in order to avoid non-identifiability. if (spaMM.getOption("example_maxtime")>7) { data("Gryphon") gry_prec <- as_precision(Gryphon_A) gry_GE <- fitmv( submodels=list(BWT ~ 1 + corrMatrix(0+mv(1,2)|ID)+(0+mv(1,2)|ID), TARSUS ~ 1 + corrMatrix(0+mv(1,2)|ID)+(0+mv(1,2)|ID)), fixed=list(phi=c(1e-6,1e-6)), corrMatrix = gry_prec, data = Gryphon_df, method = "REML") # Estimates are practically identical to those reported for package # 'asreml' (https://www.vsni.co.uk/software/asreml-r) # according to Supplementary File 3 of Wilson et al., p.7: lambda_table <- summary(gry_GE, digits=5,verbose=FALSE)$lambda_table by_spaMM <- na.omit(unlist(lambda_table[,c("Var.","Corr.")]))[1:6] by_asreml <- c(3.368449,12.346304,3.849875,17.646017,0.381463,0.401968) by_spaMM/by_asreml-1 # relative differences ~ O(1e-4) }
#### Bivariate-response model used as example in Wilson et al. (2010): # joint modelling of birth weight (BWT) and tarsus length (TARSUS). # The relatedness matrix is specified as a 'corrMatrix'. The random # effect 'corrMatrix(0+mv(1,2)|ID)' then represents genetic effects # correlated over traits and individuals (see help("composite-ranef")). # The ...(0+...) syntax avoids contrasts being used in the design # matrix of the random effects, as it would not does make much sense # to represent TARSUS as a contrast to BWT. # The relatedness matrix will be specified through its inverse, # using as_precision(), so that spaMM does not have to find out and # inform the user that using the inverse is better (as is typically # the case for relatedness matrices). But using as_precision() is # not required. See help("algebra") for Details. # The second random effect '(0+mv(1,2)|ID)' represents correlated # environmental effects. Since measurements are not repeated within # individuals, this effect also absorbs all residual variation. The # residual variances 'phi' must then be fixed to some negligible values # in order to avoid non-identifiability. if (spaMM.getOption("example_maxtime")>7) { data("Gryphon") gry_prec <- as_precision(Gryphon_A) gry_GE <- fitmv( submodels=list(BWT ~ 1 + corrMatrix(0+mv(1,2)|ID)+(0+mv(1,2)|ID), TARSUS ~ 1 + corrMatrix(0+mv(1,2)|ID)+(0+mv(1,2)|ID)), fixed=list(phi=c(1e-6,1e-6)), corrMatrix = gry_prec, data = Gryphon_df, method = "REML") # Estimates are practically identical to those reported for package # 'asreml' (https://www.vsni.co.uk/software/asreml-r) # according to Supplementary File 3 of Wilson et al., p.7: lambda_table <- summary(gry_GE, digits=5,verbose=FALSE)$lambda_table by_spaMM <- na.omit(unlist(lambda_table[,c("Var.","Corr.")]))[1:6] by_asreml <- c(3.368449,12.346304,3.849875,17.646017,0.381463,0.401968) by_spaMM/by_asreml-1 # relative differences ~ O(1e-4) }
This gets “leverages” or “hat values” from an object. However, there is hidden complexity in what this may mean, so care must be used in selecting proper arguments for a given use (see Details). To get the full hat matrix, see get_matrix(., which="hat_matrix")
.
## S3 method for class 'HLfit' hatvalues(model, type = "projection", which = "resid", force=FALSE, ...)
## S3 method for class 'HLfit' hatvalues(model, type = "projection", which = "resid", force=FALSE, ...)
model |
An object of class |
type |
Character: |
which |
Character: |
force |
Boolean: to force recomputation of the leverages even if they are available in the object, for checking purposes. |
... |
For consistency with the generic. |
Leverages may have distinct meaning depending on context. The textbook version for linear models is that leverages are the diagonal elements of a projection matrix (“hat matrix”), and that they may be used to standardize (“studentize”) residuals as follows. If the residual variance
is known, then the variance of each fitted residual
is
. Standardized residuals, all with variance 1, are then
. This standardization of variance no longer holds exactly with estimated
, but if one uses here an unbiased (REML) estimator of
, the studentized residuals may still practically have a unit expected variance.
When a simple linear model is fitted by ML, the variance of the fitted residuals is less than , but
is downward biased so that residuals standardized only by
, without any leverage correction, more closely have expected unit variance than if corrected by the previous leverages. The ML and REML computations can be seen as both using “standardizing” leverages, defined so that they are zero in the ML case and are equal to the “projection” leverages (the above ones, derived from a projection matrix) in the REML case.
These “standardizing” leverages can themselves been seen as special cases of those that appear in expressions for derivatives, with respect to the dispersion parameters, of the log-determinant of the information matrices considered in the Laplace approximation for marginal or restricted likelihood (Lee et al. 2006). This provides a basis to generalize the concept of standardizing leverages for ML and REML in mixed-effect models. In particular, in an ML fit, one considers leverages that are no longer the diagonal elements of the projection matrix for the mixed model [and, as hinted above, for a simple linear model the ML
are zero]. The generalized standardizing leverages may include corrections for non-Gaussian response, for non-Gaussian random effects, and for taking into account the variation of the GLM weights in the logdet(info.mat) derivatives. Which corrections are included depend on the precise method used to fit the model (e.g., EQL vs PQL vs REML). Standardizing leverages are also defined for the random effects.
These distinctions suggest breaking the usual synonymy between “leverages” or “hat values”: the term “hat values” better stands for the diagonal elements of a projection matrix, while “leverages” better stands for the standardizing values.
hatvalues(.,type="std")
returns the standardizing leverages. By contrast, hatvalues(.,type="projection")
will always return hat values from the fitted projection matrix. Note that these values typically differ between ML and REML fit because the fitted projection matrix differs between them.
A list with separate components resid
(leverages of the observations) and ranef
if which="both"
, and a vector otherwise.
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006) Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
if (spaMM.getOption("example_maxtime")>0.8) { data("Orthodont",package = "nlme") rnge <- (107:108) # all different: # hatvalues(rlfit <- fitme(distance ~ age+(age|Subject), data = Orthodont, method="REML"))[rnge] hatvalues(mlfit <- fitme(distance ~ age+(age|Subject), data = Orthodont))[rnge] hatvalues(mlfit,type="std")[rnge] }
if (spaMM.getOption("example_maxtime")>0.8) { data("Orthodont",package = "nlme") rnge <- (107:108) # all different: # hatvalues(rlfit <- fitme(distance ~ age+(age|Subject), data = Orthodont, method="REML"))[rnge] hatvalues(mlfit <- fitme(distance ~ age+(age|Subject), data = Orthodont))[rnge] hatvalues(mlfit,type="std")[rnge] }
A fitting function acting as a convenient interface for HLfit
, constructing the correlation matrix of random effects from the arguments, then estimating fixed effects and dispersion parameters using HLfit
. Various arguments are available to constrain the correlation structure, covStruct
and distMatrix
being the more general ones (for any number of random effects), and adjMatrix
and corrMatrix
being alternatives to covStruct
for a single correlated random effect. uniqueGeo
is deprecated.
HLCor(formula, data, family = gaussian(), fixed=NULL, ranPars, distMatrix, adjMatrix, corrMatrix, covStruct=NULL, method = "REML", verbose = c(inner=FALSE), control.dist = list(), weights.form = NULL, ...)
HLCor(formula, data, family = gaussian(), fixed=NULL, ranPars, distMatrix, adjMatrix, corrMatrix, covStruct=NULL, method = "REML", verbose = c(inner=FALSE), control.dist = list(), weights.form = NULL, ...)
formula |
A |
fixed , ranPars
|
A list of given values for correlation parameters (some of which are mandatory), and possibly also dispersion parameters
(optional, but passed to HLfit if present). |
data |
The data frame to be analyzed. |
family |
A |
distMatrix |
Either a distance matrix between geographic locations, forwarded to |
adjMatrix |
An single adjacency matrix, used if a random effect of the form |
corrMatrix |
A matrix C used if a random effect term of the form |
covStruct |
An interface for specifying correlation structures for different types of random effect ( |
method |
Character: the fitting method to be used, such as |
weights.form |
Specification of prior weights by a one-sided formula: use |
verbose |
A vector of booleans. |
control.dist |
A list of arguments that control the computation of the distance argument of the correlation functions. Possible elements are
|
... |
Further parameters passed to |
For approximations of likelihood, see method
. For the possible structures of random effects, see random-effects
, but note that HLCor
cannot adjust parameters of correlation models (with the exception of conditional autoregressive ones). Any such parameter must be specified by the ranPars
argument. More generally, the correlation matrix for random effects can be specified by various combinations of formula terms and other arguments (see Examples):
Matern(1|<...>)
, using the spatial coordinates in <...>
. This will construct a correlation matrix according to the Matérn correlation function (see MaternCorr
);
Cauchy(1|<...>)
, as for Matern (see CauchyCorr
);
as provided by distMatrix
(see Examples);
corrMatrix(1|<...>)
with corrMatrix
argument. See corrMatrix
for further details.
adjacency(1|<...>)
with adjMatrix
. See adjacency
for further details;
AR1(1|<...>)
See AR1
for further details.
The return value of an HLfit
call, with the following additional attributes:
HLCorcall |
the HLCor call |
info.uniqueGeo |
Unique geographic locations. |
autoregressive
for additional examples, MaternCorr
, HLfit
, and corrHLfit
# Example with an adjacency matrix (autoregressive model): # see 'adjacency' documentation page #### Matern correlation using only the Matern() syntax data("blackcap") (fitM <- HLCor(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML", ranPars=list(nu=0.6285603,rho=0.0544659))) #### Using the 'distMatrix' argument data("blackcap") # # Build distance matrix (here equivalent to the default one for a Matern() term) MLdistMat <- as.matrix(proxy::dist(blackcap[,c("latitude","longitude")])) # (fitD <- HLCor(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, distMatrix=MLdistMat, method="ML", ranPars=list(nu=0.6285603,rho=0.0544659))) # : result here must be equivalent to the one without the distMatrix. diff(c(logLik(fitM),logLik(fitD)))
# Example with an adjacency matrix (autoregressive model): # see 'adjacency' documentation page #### Matern correlation using only the Matern() syntax data("blackcap") (fitM <- HLCor(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, method="ML", ranPars=list(nu=0.6285603,rho=0.0544659))) #### Using the 'distMatrix' argument data("blackcap") # # Build distance matrix (here equivalent to the default one for a Matern() term) MLdistMat <- as.matrix(proxy::dist(blackcap[,c("latitude","longitude")])) # (fitD <- HLCor(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, distMatrix=MLdistMat, method="ML", ranPars=list(nu=0.6285603,rho=0.0544659))) # : result here must be equivalent to the one without the distMatrix. diff(c(logLik(fitM),logLik(fitD)))
This function fits GL(M)Ms as well as some hierarchical generalized linear models (HGLM; Lee and Nelder 2001). It may be called on its own but is now better seen as a backend for the main fitting function fitme
(or fitmv
for multivariate-response models). This documentation completes the documentation of the latter functions with respect to some arguments they pass to HLfit
and with respect to the structure of the objects they return.
On its own, HLfit
fits both fixed effects parameters, and dispersion parameters i.e. the variance of the random effects (full covariance for random-coefficient models), and the variance of the residual error. The linear predictor is of the standard form offset+ X beta + Z b
, where X is the design matrix of fixed effects and Z is a design matrix of random effects (typically an incidence matrix with 0s and 1s, but not necessarily). Models are fitted by an iterative algorithm alternating estimation of fixed effects and of dispersion parameters. The residual dispersion may follow a “structured-dispersion model” modeling heteroscedasticity.
Estimation of the latter parameters is performed by a form of fit of debiased residuals, which allows fitting a structured-dispersion model (Smyth et al. 2001). However, evaluation of the debiased residuals can be slow in particular for large datasets. For models without structured dispersion, it is then worth using the fitme
function. Ths function (as well as corrHLfit
) can optimize the likelihood of HLfit
fits for different given values of the dispersion parameters (“outer optimization”), thereby avoiding the need to estimate debiased residuals.
HLfit(formula, data, family = gaussian(), rand.family = gaussian(), resid.model = ~1, REMLformula = NULL, verbose = c(inner = FALSE), HLmethod = "HL(1,1)", method="REML", control.HLfit = list(), control.glm = list(), init.HLfit = list(), fixed=list(), ranFix, etaFix = list(), prior.weights = NULL, weights.form = NULL, processed = NULL) ## see 'rand.family' argument for inverse.Gamma
HLfit(formula, data, family = gaussian(), rand.family = gaussian(), resid.model = ~1, REMLformula = NULL, verbose = c(inner = FALSE), HLmethod = "HL(1,1)", method="REML", control.HLfit = list(), control.glm = list(), init.HLfit = list(), fixed=list(), ranFix, etaFix = list(), prior.weights = NULL, weights.form = NULL, processed = NULL) ## see 'rand.family' argument for inverse.Gamma
formula |
A |
data |
A data frame containing the variables named in the model formula. |
family |
A |
rand.family |
A |
resid.model |
Used to specify a model for the dispersion parameter of the mean-response family. See the |
REMLformula |
A model |
verbose |
A vector of booleans or integers. The |
method |
Character: the fitting method.
allowed values include |
HLmethod |
Same as |
control.HLfit |
A list of parameters controlling the fitting algorithms, which should mostly be ignored in routine use.
See |
control.glm |
List of parameters controlling calls to |
init.HLfit |
A list of initial values for the iterative algorithm, with possible elements of the list are
|
fixed , ranFix
|
A list of fixed values of random effect parameters. |
etaFix |
A list of given values of the coefficients of the linear predictor. See |
prior.weights |
An optional vector of prior weights as in |
weights.form |
Specification of prior weights by a one-sided formula: use |
processed |
A list of preprocessed arguments, for programming purposes only. |
I. Approximations of likelihood: see method
.
II. Possible structure of Random effects: see random-effects
, but note that HLfit
does not fit models with autocorrelated random effects.
III. The standard errors reported may sometimes be misleading. For each set of parameters among ,
, and
parameters these are computed assuming that the other parameters are known without error. This is why they are labelled
Cond. SE
(conditional standard error). This is most uninformative in the unusual case where and
are not separately estimable parameters. Further, the SEs for
and
are rough approximations as discussed in particular by Smyth et al. (2001;
method).
IV. prior weights. This controls the likelihood analysis of heteroscedastic models. In particular, changing the weights by a constant factor f should, and will, yield a fit with unchanged likelihood and (Intercept) estimates of phi
also increased by f (except if a non-trivial resid.formula
with log link is used). This is consistent with what glm
does, but other packages may not follow this logic (whatever their documentation may say: check by yourself by changing the weights by a constant factor). Further, post-fit functiosn (in particular those extracting various forms of residuals) may be inconsistent in their handling of prior weights.
An object of class HLfit
, which is a list with many elements, not all of which are documented.
Various extractor functions are available (see extractors
, vcov
, get_fittedPars
, get_matrix
, and so on). They should be used as far as possible as they should be backward-compatible from version 2.0.0 onwards, while the structure of the return object may still evolve. The following information may be useful for extracting further elements of the object.
Elements include descriptors of the fit:
eta |
Fitted values on the linear scale (including the predicted random effects). |
fv |
Fitted values ( |
fixef |
The fixed effects coefficients, |
v_h |
The random effects on the linear scale, |
phi |
The residual variance |
phi.object |
A possibly more complex object describing |
lambda |
The random-effect ( |
lambda.object |
A possibly more complex object describing |
ranef_info |
environment where information about the structure of random effects is stored (see |
corrPars |
Agglomerates information on correlation parameters, either fixed, or estimated ((see |
APHLs |
A list whose elements are various likelihood components, including conditional likelihood, h-likelihood, and the Laplace approximations: the (approximate) marginal likelihood |
The covariance matrix of estimates is not included as such, but can be extracted by
vcov
.
Information about the input is contained in output elements named as arguments of the fitting function calls (data,family,resid.family,ranFix,prior.weights
), with the following notable exceptions or modifications:
predictor |
The |
resid.predictor |
Analogous to |
rand.families |
corresponding to the |
Further miscellaneous diagnostics and descriptors of model structure:
X.pv |
The design matrix for fixed effects (returned by the |
ZAlist , strucList
|
Two lists of matrices, respectively the design matrices “Z”, and the “L” matrices, for the different random-effect terms. The extractor |
BinomialDen |
(binomial data only) the binomial denominators; |
y |
the response vector; for binomial data, the frequency response. |
models |
Additional information on model structure for |
HL |
A set of indices that characterize the approximations used for likelihood; |
leve_phi , lev_lambda
|
Leverages (see |
dfs |
list (possibly structured): some information about degrees of freedom for different components of the model. But its details may be difficult to interpret and the |
how |
A list containing the information properly extracted by the |
warnings |
A list of warnings for events that may have occurred during the fit. |
Finally, the object includes programming tools: call, spaMM.version, fit_time
and an environment envir
that may contain whatever may be needed in some post-fit operations..
Lee, Y., Nelder, J. A. (2001) Hierarchical generalised linear models: A synthesis of generalised linear models, random-effect models and structured dispersions. Biometrika 88, 987-1006.
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006). Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
Smyth GK, Huele AF, Verbyla AP (2001). Exact and approximate REML for heteroscedastic regression. Statistical Modelling 1, 161-175.
HLCor
for estimation with given spatial correlation parameters;
corrHLfit
for joint estimation with spatial correlation parameters;
fitme
as an alternative to all these functions.
data("wafers") ## Gamma GLMM with log link HLfit(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) ## Gamma - inverseGamma HGLM with log link HLfit(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), rand.family=inverse.Gamma(log), resid.model = ~ X3+I(X3^2) , data=wafers)
data("wafers") ## Gamma GLMM with log link HLfit(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) ## Gamma - inverseGamma HGLM with log link HLfit(y ~ X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), family=Gamma(log), rand.family=inverse.Gamma(log), resid.model = ~ X3+I(X3^2) , data=wafers)
how
is defined as a generic with currently only one non-default method, for objects of class HLfit
. This method provide information about how such a fit was obtained.
how(object, ...) ## S3 method for class 'HLfit' how(object, devel=FALSE, verbose=TRUE, format=print, ...) ## S3 method for class 'HLfitlist' how(object, devel=FALSE, verbose=TRUE, format=print, ...)
how(object, ...) ## S3 method for class 'HLfit' how(object, devel=FALSE, verbose=TRUE, format=print, ...) ## S3 method for class 'HLfitlist' how(object, devel=FALSE, verbose=TRUE, format=print, ...)
object |
Any R object. |
devel |
Boolean; Whether to provide additional cryptic information. For development purposes, not further documented. |
verbose |
Boolean; Whether to print information about the input object. |
format |
wrapper for printing format. E.g., |
... |
Other arguments that may be needed by some method. |
A list, returned invisibly, whose elements are not further described here, some being slightly cryptic or subject to future changes However, how(.)$fit_time
is a clean way of getting the fit time. If verbose
is TRUE
, the function prints a message presenting some of these elements.
foo <- HLfit(y~x, data=data.frame(x=runif(3), y=runif(3)), method="ML", ranFix=list(phi=1)) how(foo)
foo <- HLfit(y~x, data=data.frame(x=runif(3), y=runif(3)), method="ML", ranFix=list(phi=1)) how(foo)
Several parameters (notably the dispersion parameters: the variance of random effects and the residual variance parameter, if any) can be estimated either by iterative algorithms, or by generic optimization methods. The development of the fitme
function aims to provide full control of the selection of algorithms. For example, if two random effects are fitted, theninit=list(lambda=c(NA,NaN))
enforces generic optimization for the first variance and iterative algorithms for the second.init=list(lambda=c(0.1,NaN))
has the same effect and additionally provides control of the initial value for optimization (whereas init.HLfit=list(lambda=c(NA,0.1))
will provide control of the initial value for iterations).
How to know which algorithm has been selected for each parameter? fitme(., verbose=c(TRACE=TRUE))
shows successive values of the variables estimated by optimization (See Examples; if no value appears, then all are estimated by iterative methods). The first lines of the summary of a fit object should tell which variances are estimated by the “outer” method.
corrHLfit
, which uses inner optimization by default, can be forced to perform outer optimization. Its control is more limited, as NA
s and NaN
s are not allowed. Instead, only numeric values as in init=list(lambda=0.1)
are allowed.
## Not run: air <- data.frame(passengers = as.numeric(AirPassengers), year_z = scale(rep(1949:1960, each = 12)), month = factor(rep(1:12, 12))) air$time <- 1:nrow(air) # Use verbose to find that lambda is estimated by optimization fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE)) # Use init to enforce iterative algorithm for lambda estimation: fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE), init=list(lambda=NaN)) # (but then it may be better to enforce it also for phi: init=list(lambda=NaN, phi=NaN)) # # Use init to enforce generic optimization for lambda estimation, # and control initial value: fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE), init=list(lambda=0.1)) # See help("multinomial") for more examples of control by initial values. ## End(Not run)
## Not run: air <- data.frame(passengers = as.numeric(AirPassengers), year_z = scale(rep(1949:1960, each = 12)), month = factor(rep(1:12, 12))) air$time <- 1:nrow(air) # Use verbose to find that lambda is estimated by optimization fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE)) # Use init to enforce iterative algorithm for lambda estimation: fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE), init=list(lambda=NaN)) # (but then it may be better to enforce it also for phi: init=list(lambda=NaN, phi=NaN)) # # Use init to enforce generic optimization for lambda estimation, # and control initial value: fitme(passengers ~ month * year_z + AR1(1|time), data = air, verbose=c(TRACE=TRUE), init=list(lambda=0.1)) # See help("multinomial") for more examples of control by initial values. ## End(Not run)
For dispersion parameter ,
Gamma
means that random effects are distributed as Gamma(
shape=
1/,scale=
), so
has mean 1 and variance
. Both the log (
) and identity (
) links are possible, though in the latter case the variance of
is constrained below 1 (otherwise Laplace approximations fail).
The two-parameter inverse Gamma distribution is the distribution of the reciprocal of a variable distributed according to the Gamma distribution Gamma with the same shape and scale parameters. inverse.Gamma
implements the one-parameter inverse Gamma family with shape=1+1/ and rate=1/
) (rate=1/scale). It is used to model the distribution of random effects. Its mean=1; and its variance =
if
, otherwise infinite. The default link is
"-1/mu"
, in which case v=-1/u
is “-Gamma”-distributed with the same shape and rate, hence with mean and variance
, which is a different one-parameter Gamma family than the above-described
Gamma
. The other possible link is v=log(u)
in which caseGamma
, with mean
digamma
and variance trigamma(
).
inverse.Gamma(link = "-1/mu") # Gamma(link = "inverse") using stats::Gamma
inverse.Gamma(link = "-1/mu") # Gamma(link = "inverse") using stats::Gamma
link |
For |
# see help("HLfit") for fits using the inverse.Gamma distribution.
# see help("HLfit") for fits using the inverse.Gamma distribution.
Separation occurs in binomial response models when a combination of the predictor variables perfectly predict a level of the response. In such a case the estimates of the coefficients for these variables diverge to (+/-)infinity, and the numerical algorithms typically fail. To anticipate such a problem, the fitting functions in spaMM
try to check for separation by default. The check may take much time, and is skipped if the “problem size” exceeds a threshold defined by spaMM.options(separation_max=<.>)
, in which case a message will tell users by how much they should increase separation_max
to force the check (its exact meaning and default value are subject to changes without notice but the default value aims to correspond to a separation check time of the order of 1s on the author's computer).
is_separated
is a convenient interface to procedures from the ROI
package, allowing them to be called explicitly by the user to check bootstrap samples (see Example in anova
).
is_separated.formula
is a variant (not yet a formal S3 method) that performs the same check, but using arguments similar to those of fitme(., family=binomial())
.
is_separated(x, y, verbose = TRUE, solver=spaMM.getOption("sep_solver")) is_separated.formula(formula, ..., separation_max=spaMM.getOption("separation_max"), solver=spaMM.getOption("sep_solver"))
is_separated(x, y, verbose = TRUE, solver=spaMM.getOption("sep_solver")) is_separated.formula(formula, ..., separation_max=spaMM.getOption("separation_max"), solver=spaMM.getOption("sep_solver"))
x |
Design matrix for fixed effects. |
y |
Numeric response vector |
formula |
A model formula |
... |
|
separation_max |
numeric: non-default value allow for easier local control of this spaMM option. |
solver |
character: name of linear programming solver used to assess separation; passed to |
verbose |
Whether to print some messages (e.g., pointing model terms that cause separation) or not. |
Returns a boolean; TRUE
means there is (quasi-)separation. Screen output may give further information, such as pointing model terms that cause separation.
The method accessible by solver="glpk"
implements algorithms described by
Konis, K. 2007. Linear Programming Algorithms for Detecting Separated Data in Binary Logistic Regression Models. DPhil Thesis, Univ. Oxford. https://ora.ox.ac.uk/objects/uuid:8f9ee0d0-d78e-4101-9ab4-f9cbceed2a2a.
See also the 'safeBinaryRegression' and 'detectseparation' package.
set.seed(123) d <- data.frame(success = rbinom(10, size = 1, prob = 0.9), x = 1:10) is_separated.formula(formula= success~x, data=d) # FALSE is_separated.formula(formula= success~I(success^2), data=d) # TRUE
set.seed(123) d <- data.frame(success = rbinom(10, size = 1, prob = 0.9), x = 1:10) is_separated.formula(formula= success~x, data=d) # FALSE is_separated.formula(formula= success~I(success^2), data=d) # TRUE
A data set from Tonnabel et al. (2021) to be fitted by models with sex-specific spatial random effects. Leucadrendron rubrum is a dioecious shrub from South Africa. Various phenotypes were recorded on individuals from a small patch of habitat.
data("Leuca")
data("Leuca")
Leuca
is
'data.frame': 156 obs. of 12 variables: $ name : Factor w/ 156 levels "f_101","f_102",..: 1 2 3 4 5 6 7 8 9 10 ... $ sex : Factor w/ 2 levels "f","m": 1 1 1 1 1 1 1 1 1 1 ... $ area : num 0.857 0.9 0.827 0.654 0.733 ... $ diam : int 60 30 180 50 70 80 130 90 27 59 ... $ fec : num 0.013 0.0137 5.1171 0.2905 1.042 ... $ fec_div: num 0.0128 0.0135 5.037 0.2859 1.0257 ... $ x : num 42 41 62.5 58.5 42.5 33.5 24 26.5 25 41 ... $ y : num 23 46 58 63 51 51 55.5 55.5 58.5 63 ... $ diamZ : num -0.713 -1.479 2.352 -0.968 -0.457 ... $ areaZ : num 0.72 0.92 0.586 -0.2 0.158 ... $ male : logi FALSE FALSE FALSE FALSE FALSE FALSE ... $ female : logi TRUE TRUE TRUE TRUE TRUE TRUE ...
Tonnabel, J., Klein, E.K., Ronce, O., Oddou-Muratorio, S., Rousset, F., Olivieri, I., Courtiol, A. and Mignot, A. (2021), Sex-specific spatial variation in fitness in the highly dimorphic Leucadendron rubrum. Mol Ecol, 30: 1721-1735. doi:10.1111/mec.15833
MaternCorr
and composite-ranef
for examples using these data.
data(Leuca)
data(Leuca)
It may be straightforward to add columns of indicator variables for each level of a factor to the data, by
<data> <- cbind(
<data>, model.matrix( ~
<factor> - 1, data =
<data>))
.
Alternatively, indicator variables can be created on the fly for given levels, using the lev2bool
function.
lev2bool(fac, lev)
lev2bool(fac, lev)
fac |
An object coercible to |
lev |
The level of |
A one-column matrix.
Example in GxE
for alternative to using lev2bool
in specification of
random effects with heteroscedasticity, useful when the latter is controlled
by a factor with many levels.
## Elementary bivariate-response model # Data preparation # fam <- rep(c(1,2),rep(6,2)) # define two biological 'families' ID <- gl(6,2) # define 6 'individuals' resp <- as.factor(rep(c("x","y"),6)) # distinguishes two responses per individual set.seed(123) toymv <- data.frame( fam = factor(fam), ID = ID, resp = resp, y = 1 + (resp=="x") + rnorm(4)[2*(resp=="x")+fam] + rnorm(12)[6*(resp=="x")+as.integer(ID)] ) toymv <- cbind(toymv, model.matrix( ~ resp - 1, data = toymv)) # fit response-specific variances of random effect and residuals: # (fitme(y ~ resp+ (0+respx|fam)+ (0+respy|fam), resid.model = ~ 0+resp ,data=toymv)) # Same result by different syntaxes: # * by the lev2bool() specifier: (fitme(y ~ resp+ (0+lev2bool(resp,"x")|fam)+ (0+lev2bool(resp,"y")|fam), resid.model = ~ 0+resp ,data=toymv)) # * or by random-coefficient model using 'resp' factor: (fitme(y ~ resp+ (0+resp|fam), resid.model = ~ 0+resp ,data=toymv, fixed=list(ranCoefs=list("1"=c(NA,0,NA))))) # # For factors with more levels, the following function may be useful to specify # through partially fixed ranCoefs that covariances are fixed to zero: ranCoefs_for_diag <- function(nlevels) { vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } # see application in help("GxE"). # * or by the dummy() specifier from lme4: # (fitme(y ~ resp+ (0+dummy(resp,"x")|fam)+ (0+dummy(resp,"y")|fam), # resid.model = ~ 0+resp ,data=toymv))
## Elementary bivariate-response model # Data preparation # fam <- rep(c(1,2),rep(6,2)) # define two biological 'families' ID <- gl(6,2) # define 6 'individuals' resp <- as.factor(rep(c("x","y"),6)) # distinguishes two responses per individual set.seed(123) toymv <- data.frame( fam = factor(fam), ID = ID, resp = resp, y = 1 + (resp=="x") + rnorm(4)[2*(resp=="x")+fam] + rnorm(12)[6*(resp=="x")+as.integer(ID)] ) toymv <- cbind(toymv, model.matrix( ~ resp - 1, data = toymv)) # fit response-specific variances of random effect and residuals: # (fitme(y ~ resp+ (0+respx|fam)+ (0+respy|fam), resid.model = ~ 0+resp ,data=toymv)) # Same result by different syntaxes: # * by the lev2bool() specifier: (fitme(y ~ resp+ (0+lev2bool(resp,"x")|fam)+ (0+lev2bool(resp,"y")|fam), resid.model = ~ 0+resp ,data=toymv)) # * or by random-coefficient model using 'resp' factor: (fitme(y ~ resp+ (0+resp|fam), resid.model = ~ 0+resp ,data=toymv, fixed=list(ranCoefs=list("1"=c(NA,0,NA))))) # # For factors with more levels, the following function may be useful to specify # through partially fixed ranCoefs that covariances are fixed to zero: ranCoefs_for_diag <- function(nlevels) { vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } # see application in help("GxE"). # * or by the dummy() specifier from lme4: # (fitme(y ~ resp+ (0+dummy(resp,"x")|fam)+ (0+dummy(resp,"y")|fam), # resid.model = ~ 0+resp ,data=toymv))
Some “family” objects in spaMM describe models with non-GLM response families, such as the negbin1
or beta_resp
families already widely considered in previous works and other packages. These models are characterized by a linear predictor, a link function, and a distribution for residual variation that does not belong to the exponential family from which GLMs are defined.
These family objects are conceived for use with spaMM's fitting functions. They cannot generally be used as argument to the glm
function, except when this function is highjacked by use of the method="llm.fit"
argument, where llm stands for Link-Linear (as in “log-linear”, say) regression Model.
Mixed-effect models fitted by such methods cannot use expected-Hessian approximations, in contrast to GLM response families. negbin2
is a family object for a GLM response family (strictly speaking, only for fixed shape and untruncated version) but implemented as an LL-family
, in particular using only the observed Hessian matrix.
# glm(..., method="llm.fit") ## See also 'beta_resp', 'negbin1', 'betabin', and possibly later additions.
# glm(..., method="llm.fit") ## See also 'beta_resp', 'negbin1', 'betabin', and possibly later additions.
These family objects are lists, formally of class c("LLF", "family")
. Compared to a family
object, they have additional elements, not documented here.
As stats::
GLM family objects do, they provide deviance residuals through the dev.resids
member function. There are various definitions of deviance residuals for non-GLM families in the literature. Here they are defined as “2*(saturated_logLik - logLik)”, where the likelihood for the saturated model is the likelihood maximized wrt to the mean parameter for each observation
y
independently. The maximizing is not equal to the observation, in contrast to the standard result for GLMs.
data(scotlip) ### negbin1 response: # Fixed-effect model # (var_shape <- fitme(cases~I(prop.ag/10)+offset(log(expec)),family=negbin1(), data=scotlip)) # Highjacking glm(): the family parameter must be given # fitted_shape <- residVar(var_shape,which="fam_parm") glm(cases~I(prop.ag/10)+offset(log(expec)),family=negbin1(shape=fitted_shape), method="llm.fit", data=scotlip) ### Similar exercice with Beta response family: set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE)) # Fixed-effect model (var_prec <- fitme(y ~1, family=beta_resp(), data= beta_dat)) # Highjacking glm(): fitted_prec <- residVar(var_prec,which="fam_parm") glm(y ~1, family=beta_resp(prec=fitted_prec), data= beta_dat, method="llm.fit")
data(scotlip) ### negbin1 response: # Fixed-effect model # (var_shape <- fitme(cases~I(prop.ag/10)+offset(log(expec)),family=negbin1(), data=scotlip)) # Highjacking glm(): the family parameter must be given # fitted_shape <- residVar(var_shape,which="fam_parm") glm(cases~I(prop.ag/10)+offset(log(expec)),family=negbin1(shape=fitted_shape), method="llm.fit", data=scotlip) ### Similar exercice with Beta response family: set.seed(123) beta_dat <- data.frame(y=runif(100),grp=sample(2,100,replace = TRUE)) # Fixed-effect model (var_prec <- fitme(y ~1, family=beta_resp(), data= beta_dat)) # Highjacking glm(): fitted_prec <- residVar(var_prec,which="fam_parm") glm(y ~1, family=beta_resp(prec=fitted_prec), data= beta_dat, method="llm.fit")
This data set describes prevalence of infection by the nematode Loa loa in North Cameroon, 1991-2001. This is a superset of the data discussed by Diggle and Ribeiro (2007) and Diggle et al. (2007). The study investigated the relationship between altitude, vegetation indices, and prevalence of the parasite.
data("Loaloa")
data("Loaloa")
The data frame includes 197 observations on the following variables:
latitude, in degrees.
longitude, in degrees.
sample size per location
number of infected individuals per location
maximum normalised-difference vegetation index (NDVI) from repeated satellite scans
standard error of NDVI
altitude, in m.
Additional altitude variables derived from the previous one, provided for convenience: respectively, positive values of altitude-650, positive values of altitude-1000, and positive values of altitude-1300
a copy of maxNDVI modified as maxNDVI1[maxNDVI1>0.8] <- 0.8
The data were last retrieved on March 1, 2013 from P.J. Ribeiro's web resources
atwww.leg.ufpr.br/doku.php/pessoais:paulojus:mbgbook:datasets
. A current (2022-06-18) source is
https://www.lancaster.ac.uk/staff/diggle/moredata/Loaloa.txt).
Diggle, P., and Ribeiro, P. 2007. Model-based geostatistics, Springer series in statistics, Springer, New York.
Diggle, P. J., Thomson, M. C., Christensen, O. F., Rowlingson, B., Obsomer, V., Gardon, J., Wanji, S., Takougang, I., Enyong, P., Kamgno, J., Remme, J. H., Boussinesq, M., and Molyneux, D. H. 2007. Spatial modelling and the prediction of Loa loa risk: decision making under uncertainty, Ann. Trop. Med. Parasitol. 101, 499-509.
data("Loaloa") if (spaMM.getOption("example_maxtime")>5) { fitme(cbind(npos,ntot-npos)~1 +Matern(1|longitude+latitude), data=Loaloa, family=binomial()) } ### Variations on the model fit by Diggle et al. ### on a subset of the Loaloa data ### In each case this shows the slight differences in syntax, ### and the difference in 'typical' computation times, ### when fit using corrHLfit() or fitme(). if (spaMM.getOption("example_maxtime")>4) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="HL(0,1)", data=Loaloa,family=binomial(),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>1.6) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="HL(0,1)", data=Loaloa,family=binomial(),fixed=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>5.8) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>2.5) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),fixed=list(nu=0.5),method="REML") } ## Diggle and Ribeiro (2007) assumed (in this package notation) Nugget=2/7: if (spaMM.getOption("example_maxtime")>7) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranFix=list(nu=0.5,Nugget=2/7)) } if (spaMM.getOption("example_maxtime")>1.3) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="REML", data=Loaloa,family=binomial(),fixed=list(nu=0.5,Nugget=2/7)) } ## with nugget estimation: if (spaMM.getOption("example_maxtime")>17) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(), init.corrHLfit=list(Nugget=0.1),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>5.5) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),method="REML", init=list(Nugget=0.1),fixed=list(nu=0.5)) }
data("Loaloa") if (spaMM.getOption("example_maxtime")>5) { fitme(cbind(npos,ntot-npos)~1 +Matern(1|longitude+latitude), data=Loaloa, family=binomial()) } ### Variations on the model fit by Diggle et al. ### on a subset of the Loaloa data ### In each case this shows the slight differences in syntax, ### and the difference in 'typical' computation times, ### when fit using corrHLfit() or fitme(). if (spaMM.getOption("example_maxtime")>4) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="HL(0,1)", data=Loaloa,family=binomial(),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>1.6) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="HL(0,1)", data=Loaloa,family=binomial(),fixed=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>5.8) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>2.5) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),fixed=list(nu=0.5),method="REML") } ## Diggle and Ribeiro (2007) assumed (in this package notation) Nugget=2/7: if (spaMM.getOption("example_maxtime")>7) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),ranFix=list(nu=0.5,Nugget=2/7)) } if (spaMM.getOption("example_maxtime")>1.3) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude),method="REML", data=Loaloa,family=binomial(),fixed=list(nu=0.5,Nugget=2/7)) } ## with nugget estimation: if (spaMM.getOption("example_maxtime")>17) { corrHLfit(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(), init.corrHLfit=list(Nugget=0.1),ranFix=list(nu=0.5)) } if (spaMM.getOption("example_maxtime")>5.5) { fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), data=Loaloa,family=binomial(),method="REML", init=list(Nugget=0.1),fixed=list(nu=0.5)) }
The anova
method for fit objects from spaMM has two uses: if a single fit object is provided, ANOVA tables may be returned, with specific procedures for univariate-response LMs, GLMs and LMMs (see Details). Alternatively, if a second fit object is provided (object2
argument), anova
performs as an alias for LRT
. The LRT
function here differs from fixedLRT
by its arguments (model fits for LRT
, but all arguments required to fit the models for fixedLRT
), and by the format of its return value.
LRT
performs a likelihood ratio (LR) test between two model fits, the “full” and the “null” model fits. It determines which model is the more complete one by comparing model components including the fixed-effect, random-effect, residual-dispersion model specifications, and response families (offsets are ignored). Then, a standard test based on the asymptotic chi-square distribution is performed. In addition, parametric bootstrap p-values can be computed, either using the raw bootstrap distribution of the likelihood ratio, or a bootstrap estimate of the Bartlett correction of the LR statistic.
These different tests perform diffferently depending on the differences between the two models:
* If the models differ only by their fixed effects, the asymptotic LRT may be anticonservative, but the Bartlett-corrected one is generally well-calibrated.
* If the two models differ by their random effects, tests based on the chi-square distribution (including their Bartlett-corrected version) may be poorly behaved, as such tests assume unbounded parameters, as contrasted to, e.g., necessarily positive variances.
In such cases the raw boostrap test may be the only reliable test. The procedure aims to detect and report such issues, but may not report all problems: users remain responsible for applying the tests in valid conditions (see Caveats in Details section). In simple cases (such as comparing a fixed-effect to a mixed-effect model with the same fixed-effect term), the chi-square tests may not be reported. In other cases (see Examples) they may otherwise be reported, with a warning when the procedure detects some cases of estimates at the boundary for the full model, or detects cases where the LR statistic of bootstrap replicates is often zero (also suggesting that estimates are at the boundary in such replicates).
* If the fits differ by the fixed effects terms of their residual-dispersion models (but not by any random effect specification), tests based on the chi-square distribution are reported. A bootstrap can be performed as in other cases.
* Tests for some cases of nested response families (e.g., the Poisson versus its extensions) are tentatively allowed.
* In some cases the full and the null models cannot be identified and the basic LRT based on the chi-square distribution will not be returned, but a bootstrap test may still be performed.
* The case where residual-dispersion models of either fit include random effects is problematic as, for such fits, the fitting procedure does not maximize the reported likelihood. The basic LRT is not returned when the two fits differ by their random effects, but is still performed otherwise (see Examples); and a bootstrap test may still be performed in both cases.
## S3 method for class 'HLfit' anova(object, object2, type = "2", method="", ...) # LRT(object, object2, boot.repl = 0L, resp_testfn = NULL, simuland = eval_replicate, # many further arguments can be passed to spaMM_boot via the '...' # These include arguments for parallel computations, such as # nb_cores, fit_env, # as well as other named arguments and spaMM_boot's own '...' ...)
## S3 method for class 'HLfit' anova(object, object2, type = "2", method="", ...) # LRT(object, object2, boot.repl = 0L, resp_testfn = NULL, simuland = eval_replicate, # many further arguments can be passed to spaMM_boot via the '...' # These include arguments for parallel computations, such as # nb_cores, fit_env, # as well as other named arguments and spaMM_boot's own '...' ...)
object |
Fit object returned by a spaMM fitting function. |
object2 |
Optional second model fit to be compared to the first (their order does not matter, except in non-standard cases where the second model is taken as teh null one and a message is issued). |
type |
ANOVA type for LMMs, as interpreted by lmerTest. Note that the default (single-term deletion ANOVA) differs from that of lmerTest. |
boot.repl |
the number of bootstrap replicates. |
resp_testfn |
See argument |
simuland |
a function, passed to |
method |
Only non-default value is |
... |
Further arguments, passed to |
* The ANOVA-table functionality has been included here mainly to provide access to F tests (including, for LMMs, the “Satterthwaite method” as developed by Fai and Cornelius, 1996), using pre-existing procedures as template or backend for expediency and familiarity:
ANOVA tables for LMs and GLMs have been conceived to replicate the functionality, output format and details of base R anova
, and therefore replicate some of their limitations, e.g., they only perform sequential analysis (“type 1”) in the same way as anova.lm
and anova.glm
. However, a difference occurs for Gamma GLMs, because the dispersion estimates for Gamma GLMs differ between stats::glm
and spaMM fits (see Details in method
). Therefore, F tests and Mallows' Cp differ too; results from spaMM REML fits being closer than ML fits to those from glm()
fits;
For LMMs, ANOVA tables are provided by interfacing lmerTest::anova
(with non-default type
). This procedure should handle all types of LMMs that can be fitted by spaMM; yet, the displayed information should be inspected to check that some fitted random-effect parameters are not ignored when computing information for the Satterthwaite method.
For fitted models that do not lay within previous categories, such as GLMMs, models with a residual-dispersion submodel, and multivariate-response models, a table of tests for single-term deletions using the classical “Wald” chi-squared test based on coefficient values and their conditional standard error estimates will be returned. LRTs (moreover, with bootstrap correction) are more reliable than such tests and, as calling them requires a second model to be explicitly specified, they may also help users thinking about the hypothesis they are testing.
* Bootstrap LRTs: A raw bootstrap p-value can be computed from the simulated distribution as (1+sum(t >= t0))/(N+1)
where t0
is the original likelihood ratio, t
the vector of bootstrap replicates and N
its length. See Davison & Hinkley (1997, p. 141) for discussion of the adjustments in this formula. However, a computationally more economical use of the bootstrap is to provide a Bartlett correction for the likelihood ratio test in small samples. According to this correction, the mean value of the likelihood ratio statistic under the null hypothesis is computed (here estimated by a parametric bootstrap) and the original LR statistic is multiplied by
where
is the number of degrees of freedom of the test.
When models differ by their random-effect specifications, distinguishing the full from the null model is not easy. In particular, equivalent models can be specified by diverse syntaxes, so a simple textual comparison of the random-effect terms may not be enough, and model specifications that hinder such a comparison should be avoided. When differences in random effects are tested, the null distribution of the LR may include a probability mass in 1: the discussion in Details of get_RLRsim_args
applies.
* Caveats: (1) An evaluated log-likelihood ratio can be slightly negative, e.g. when a fixed-effect model is compared to a mixed one, or a spatial random effect to a block effect, if parameters of the more complete model are estimated within bounds (e.g., variance>1e-06, or Matern smoothness>0.005) designed to avoid numerical singularities, while the less complete model corresponds to a boundary case (e.g., variance=0, or smoothness=0). The bootstrap procedure tries to identify these cases and then corrects slightly negative logL ratios to 0. (2) The Bartlett correction is applicable when the true distribution of the LRT departs smoothly from the chi-square distribution, but not in cases where it has a probability mass in zero (at typically occurs in the same boundary cases).
LRT
returns an object of class fixedLRT
, actually a list with typical elements (depending on the options)
fullfit |
the HLfit object for the full model; |
nullfit |
the HLfit object for the null model; |
basicLRT |
A data frame including values of the likelihood ratio chi2 statistic, its degrees of freedom, and the p-value; |
and, if a bootstrap was performed:
rawBootLRT |
A data frame including values of the likelihood ratio chi2 statistic, its degrees of freedom, and the raw bootstrap p-value; |
BartBootLRT |
A data frame including values of the Bartlett-corrected likelihood ratio chi2 statistic, its degrees of freedom, and its p-value; |
bootInfo |
a list with the following elements:
|
When ANOVA tables are computed, the return format is that of the function called (lmerTest::anova
for LMMs) or emulated (for LMs or GLMs). For GLMs, by default no test is reported, as has been the default for anova.glm
before R 4.4.0.
Bartlett, M. S. (1937) Properties of sufficiency and statistical tests. Proceedings of the Royal Society (London) A 160: 268-282.
Davison A.C., Hinkley D.V. (1997) Bootstrap methods and their applications. Cambridge Univ. Press, Cambridge, UK.
Fai AH, Cornelius PL (1996). Approximate F-tests of multiple degree of freedom hypotheses in generalised least squares analyses of unbalanced split-plot experiments. Journal of Statistical Computation and Simulation, 54(4), 363-378. doi:10.1080/00949659608811740
See also fixedLRT
for a different interface to LRTs,get_RLRsim_args
for efficient simulation-based implementation of exact likelihood ratio tests for testing the presence of variance components,as_LMLT
for the interface to lmerTest::anova
,
and summary.HLfit
(.,details=list(<true|"Wald">))
for reporting the p-value for each t-statistic in the summary table for fixed effects, either by Student's t distribution, or by the approximation of t^2 distribution by the Chi-squared distribution (“Wald's test”).
data("wafers") ## Gamma GLMM with log link m1 <- HLfit(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers,method="ML") m2 <- update(m1,formula.= ~ . -I(X2^2)) # anova(m1,m2) # LRT ## 'anova' (Wald chi-squared tests...) for GLMM or model with a 'resid.model' anova(m1) ## ANOVA table for GLM # Gamma example, from McCullagh & Nelder (1989, pp. 300-2), as in 'glm' doc: clotting <- data.frame( u = c(5,10,15,20,30,40,60,80,100), lot1 = c(118,58,42,35,27,25,21,19,18), lot2 = c(69,35,26,21,18,16,13,12,12)) spglm <- fitme(lot1 ~ log(u), data = clotting, family = Gamma, method="REML") anova(spglm, test = "F") anova(spglm, test = "Cp") anova(spglm, test = "Chisq") anova(spglm, test = "Rao") ## ANOVA table for LMM if(requireNamespace("lmerTest", quietly=TRUE)) { lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) print(anova(lmmfit)) # => Satterthwaite method, here giving p-values # quite close to traditional t-tests given by: summary(lmmfit, details=list(p_value=TRUE)) } ## Using resp_testfn argument for bootstrap LRT: ## Not run: set.seed(1L) d <- data.frame(success = rbinom(10, size = 1, prob = 0.9), x = 1:10) xx <- cbind(1,d$x) table(d$success) m_x <- fitme(success ~ x, data = d, family = binomial()) m_0 <- fitme(success ~ 1, data = d, family = binomial()) # # Bootstrap LRTs: anova(m_x, m_0, boot.repl = 100, resp_testfn=function(y) {! is_separated(xx,as.numeric(y),verbose=FALSE)}) ## End(Not run) #### Various cases were asymptotic tests may be unreliable: set.seed(123) dat <- data.frame(g = rep(1:10, e = 10), x = (x<-rnorm(100)), y = 0.1 * x + rnorm(100)) m0 <- fitme(y ~ 1, data=dat) ## (1) Models differing both by fixed and random effects: # # (note the warning for variance at boundary): # if (spaMM.getOption("example_maxtime")>11) { m <- fitme(y ~ x + (1|g), data=dat) LRT(m,m0, boot.repl = 199L) } ## See help("get_RLRsim_args") for a fast and accurate test procedure ## (2) Models differing also by residual-dispersion models: # if (spaMM.getOption("example_maxtime")>25) { m <- fitme(y ~ x + (1|g), data=dat, resid.model= ~x) LRT(m,m0, boot.repl = 99L) } ## (3) Models differing (also) by their random-effects in resid.model: # m <- fitme(y ~ x, data=dat, resid.model= ~1+(1|g)) LRT(m,m0) # no test performed
data("wafers") ## Gamma GLMM with log link m1 <- HLfit(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers,method="ML") m2 <- update(m1,formula.= ~ . -I(X2^2)) # anova(m1,m2) # LRT ## 'anova' (Wald chi-squared tests...) for GLMM or model with a 'resid.model' anova(m1) ## ANOVA table for GLM # Gamma example, from McCullagh & Nelder (1989, pp. 300-2), as in 'glm' doc: clotting <- data.frame( u = c(5,10,15,20,30,40,60,80,100), lot1 = c(118,58,42,35,27,25,21,19,18), lot2 = c(69,35,26,21,18,16,13,12,12)) spglm <- fitme(lot1 ~ log(u), data = clotting, family = Gamma, method="REML") anova(spglm, test = "F") anova(spglm, test = "Cp") anova(spglm, test = "Chisq") anova(spglm, test = "Rao") ## ANOVA table for LMM if(requireNamespace("lmerTest", quietly=TRUE)) { lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) print(anova(lmmfit)) # => Satterthwaite method, here giving p-values # quite close to traditional t-tests given by: summary(lmmfit, details=list(p_value=TRUE)) } ## Using resp_testfn argument for bootstrap LRT: ## Not run: set.seed(1L) d <- data.frame(success = rbinom(10, size = 1, prob = 0.9), x = 1:10) xx <- cbind(1,d$x) table(d$success) m_x <- fitme(success ~ x, data = d, family = binomial()) m_0 <- fitme(success ~ 1, data = d, family = binomial()) # # Bootstrap LRTs: anova(m_x, m_0, boot.repl = 100, resp_testfn=function(y) {! is_separated(xx,as.numeric(y),verbose=FALSE)}) ## End(Not run) #### Various cases were asymptotic tests may be unreliable: set.seed(123) dat <- data.frame(g = rep(1:10, e = 10), x = (x<-rnorm(100)), y = 0.1 * x + rnorm(100)) m0 <- fitme(y ~ 1, data=dat) ## (1) Models differing both by fixed and random effects: # # (note the warning for variance at boundary): # if (spaMM.getOption("example_maxtime")>11) { m <- fitme(y ~ x + (1|g), data=dat) LRT(m,m0, boot.repl = 199L) } ## See help("get_RLRsim_args") for a fast and accurate test procedure ## (2) Models differing also by residual-dispersion models: # if (spaMM.getOption("example_maxtime")>25) { m <- fitme(y ~ x + (1|g), data=dat, resid.model= ~x) LRT(m,m0, boot.repl = 99L) } ## (3) Models differing (also) by their random-effects in resid.model: # m <- fitme(y ~ x, data=dat, resid.model= ~1+(1|g)) LRT(m,m0) # no test performed
This function computes scaled distances from whichever relevant argument it can use (see Details). The result can directly by used as input for computation of the Matérn correlation matrix. It is usually called internally by HLCor, so that users may ignore it, except if they wish to control the distance used through control.dist$method
, or the parametrization of the scaling through control.dist$rho.mapping
. control.dist$method
provide access to the distances implemented in the proxy
package, as well as to "EarthChord"
and "Earth"
methods defined in spaMM
(see Details).
make_scaled_dist(uniqueGeo, uniqueGeo2=NULL, distMatrix, rho, rho.mapping=seq_len(length(rho)), dist.method="Euclidean", return_matrix=FALSE)
make_scaled_dist(uniqueGeo, uniqueGeo2=NULL, distMatrix, rho, rho.mapping=seq_len(length(rho)), dist.method="Euclidean", return_matrix=FALSE)
uniqueGeo |
A matrix of geographical coordinates (e.g. 2 columns for latitude and longitude), without replicates of the same location. |
uniqueGeo2 |
NULL, or a second matrix of geographical coordinates, without replicates of the same location.
If NULL, scaled distances among |
distMatrix |
A distance matrix. |
rho |
A scalar or vector of positive values. Scaled distance is computed as |
rho.mapping |
A set of indices controlling which elements of the |
dist.method |
|
return_matrix |
Whether to return a |
The function uses the distMatrix
argument if provided, in which case rho must be a scalar. Vectorial rho
(i.e., different scaling of different dimensions) is feasible only by providing uniqueGeo
.
The dist.method
argument gives access to distances implemented in the proxy
package, or to user-defined ones that are made accessible to proxy
through its database. Of special interest for spatial analyses are distances computed from longitude and latitude (proxy
implements "Geodesic"
and "Chord"
distances but they do not use such coordinates: instead, they use Euclidean distance for 2D computations, i.e. Euclidean distance between points on a circle rather than on a sphere). spaMM implements two such distances: "Earth"
and "EarthChord"
, using longitude and latitude inputs in that order (see Examples). The "EarthChord"
distance is the 3D Euclidean distance “through Earth”. The "Earth"
distance is also known as the orthodromic or great-circle distance, on the Earth surface. Both distances return values in km and are based on approximating the Earth by a sphere of radius 6371.009 km.
A matrix or dist
object. If there are two input matrices, rows of the return value correspond to rows of the first matrix.
data("blackcap") ## a biologically not very meaningful, but syntactically correct example of rho.mapping fitme(migStatus ~ 1 + Matern(1|longitude+latitude+means), data=blackcap, fixed=list(nu=0.5,phi=1e-6), init=list(rho=c(1,1)), control.dist=list(rho.mapping=c(1,1,2))) ## Using orthodromic distances: # order of variables in Matern(.|longitude+latitude) matters; # Matern(1|latitude+longitude) should cause a warning fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, method="ML", fixed=list(nu=0.5,phi=1e-6), control.dist=list(dist.method="Earth"))
data("blackcap") ## a biologically not very meaningful, but syntactically correct example of rho.mapping fitme(migStatus ~ 1 + Matern(1|longitude+latitude+means), data=blackcap, fixed=list(nu=0.5,phi=1e-6), init=list(rho=c(1,1)), control.dist=list(rho.mapping=c(1,1,2))) ## Using orthodromic distances: # order of variables in Matern(.|longitude+latitude) matters; # Matern(1|latitude+longitude) should cause a warning fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, method="ML", fixed=list(nu=0.5,phi=1e-6), control.dist=list(dist.method="Earth"))
These functions provide either a map of predicted response in analyzed locations, or a predicted surface. mapMM
is a straightforward representation of the analysis of the data, while filled.mapMM
uses interpolation to cope with the fact that all predictor variables may not be known in all locations on a fine spatial grid. map_ranef
maps a single spatial random effect. These three functions takes an HLfit
object as input. mapMM
calls spaMMplot2D
, which is similar but takes a more conventional (x,y,z) input.
Using filled.mapMM
may involve questionable choices. Plotting a filled contour generally requires prediction in non-observed locations, where predictor variables used in the original data analysis may be missing. In that case, the original model formula cannot be used and an alternative model (controlled by the map.formula
argument) must be used to interpolate (not smooth) the predicted values in observed locations (these predictions still resulting from the original analysis based on predictor variables). filled.mapMM
always performs such interpolation (it does not allow one to provide values for the predictor variables). As a result (1) filled.mapMM
will be slower than a mere plotting function, since it involves the analysis of spatial data; (2) the results may have little useful meaning if the effect of the original predictor variables is not correctly represented by this interpolation step. For example, prediction by interpolation may be biased in a way analogous to prediction of temperature in non-observed locations while ignoring effect of variation in altitude in such locations. Likewise, the variance
argument of filled.mapMM
allows one only to plot the prediction variance of its own interpolator, rather than that of the input object.
map_ranef
is free of the limitations of filled.mapMM
.
spaMMplot2D(x, y, z, xrange=range(x, finite = TRUE), yrange=range(y, finite = TRUE), margin=1/20, add.map= FALSE, nlevels = 20, color.palette = spaMM.colors, map.asp=NULL, col = color.palette(length(levels) - 1), plot.title=NULL, plot.axes=NULL, decorations=NULL, key.title=NULL, key.axes=NULL, xaxs = "i", yaxs = "i", las = 1, axes = TRUE, frame.plot = axes, ...) mapMM(fitobject,Ztransf=NULL,coordinates, add.points,decorations=NULL,plot.title=NULL,plot.axes=NULL,envir=-3, ...) filled.mapMM( fitobject, Ztransf = NULL, coordinates, xrange = NULL, yrange = NULL, margin = 1/20, map.formula, phi = 1e-05, gridSteps = 41, decorations = quote(points(pred[, coordinates], cex = 1, lwd = 2)), add.map = FALSE, axes = TRUE, plot.title = NULL, plot.axes = NULL, map.asp = NULL, variance = NULL, var.contour.args = list(), smoothObject = NULL, return.="smoothObject", ...) map_ranef(fitobject, re.form, Ztransf=NULL, xrange = NULL, yrange = NULL, margin = 1/20, gridSteps = 41, decorations = quote(points(fitobject$data[, coordinates], cex = 1, lwd = 2)), add.map = FALSE, axes = TRUE, plot.title=NULL, plot.axes=NULL, map.asp = NULL, ...)
spaMMplot2D(x, y, z, xrange=range(x, finite = TRUE), yrange=range(y, finite = TRUE), margin=1/20, add.map= FALSE, nlevels = 20, color.palette = spaMM.colors, map.asp=NULL, col = color.palette(length(levels) - 1), plot.title=NULL, plot.axes=NULL, decorations=NULL, key.title=NULL, key.axes=NULL, xaxs = "i", yaxs = "i", las = 1, axes = TRUE, frame.plot = axes, ...) mapMM(fitobject,Ztransf=NULL,coordinates, add.points,decorations=NULL,plot.title=NULL,plot.axes=NULL,envir=-3, ...) filled.mapMM( fitobject, Ztransf = NULL, coordinates, xrange = NULL, yrange = NULL, margin = 1/20, map.formula, phi = 1e-05, gridSteps = 41, decorations = quote(points(pred[, coordinates], cex = 1, lwd = 2)), add.map = FALSE, axes = TRUE, plot.title = NULL, plot.axes = NULL, map.asp = NULL, variance = NULL, var.contour.args = list(), smoothObject = NULL, return.="smoothObject", ...) map_ranef(fitobject, re.form, Ztransf=NULL, xrange = NULL, yrange = NULL, margin = 1/20, gridSteps = 41, decorations = quote(points(fitobject$data[, coordinates], cex = 1, lwd = 2)), add.map = FALSE, axes = TRUE, plot.title=NULL, plot.axes=NULL, map.asp = NULL, ...)
fitobject |
The return object of a corrHLfit or fitme call. |
x , y , z
|
Three vectors of coordinates, with |
re.form |
A model formula giving the single random effect term to plot, needed only if there are several spatial random effects in the fitted model. In that case, it must be formatted as |
Ztransf |
A transformation of the predicted response, given as a function whose only required argument can be a one-column matrix. The name of this argument must be
|
coordinates |
The geographical coordinates. By default they are deduced from the model formula. For example if this formula is |
xrange |
The x range of the plot (a vector of length 2); by default defined to cover all analyzed points. |
yrange |
The y range of the plot (a vector of length 2); by default defined to cover all analyzed points. |
margin |
This controls how far (in relative terms) the plot extends beyond the x and y ranges of the analyzed points, and is overriden by explicit |
map.formula |
NULL, or a formula whose left-hand side is ignored. Provides the formula used for interpolation. If NULL, a default formula with the same spatial effect(s) as in the input |
phi |
This controls the phi value assumed in the interpolation step. Ideally |
gridSteps |
The number of levels of the grid of x and y values |
variance |
Either NULL, or the name of a component of variance of prediction by the interpolator to be plotted. Must name one of the components that can be returned by |
var.contour.args |
A list of control parameters for rendering of prediction variances. See |
add.map |
Either a boolean or an explicit expression, enclosed in |
levels |
a set of levels which are used to partition the range of z. Must be strictly increasing (and finite). Areas with z values between consecutive levels are painted with the same color. |
nlevels |
if |
color.palette |
a color palette function to be used to assign colors in the plot. |
map.asp |
the y/x aspect ratio of the 2D plot area (not of the full figure including the scale). By default, the scales for x and y are identical unless the x and y ranges are too different. Namely, the scales are identical if (plotted y range)/(plotted x range) is 1/4 < . < 4, and map.asp is 1 otherwise. |
col |
an explicit set of colors to be used in the plot. This argument overrides any palette function specification. There should be one less color than levels |
plot.title |
statements which add titles to the main plot. See Details for differences between functions. |
plot.axes |
statements which draw axes (and a box) on the main plot. See Details for differences between functions. |
decorations |
Either NULL or Additional graphic statements ( |
add.points |
Obsolete, use |
envir |
Controls the environment in which |
key.title |
statements which add titles for the plot key. |
key.axes |
statements which draw axes on the plot key. |
xaxs |
the x axis style. The default is to use internal labeling. |
yaxs |
the y axis style. The default is to use internal labeling. |
las |
the style of labeling to be used. The default is to use horizontal labeling. |
axes , frame.plot
|
logicals indicating if axes and a box should be drawn, as in plot.default. |
smoothObject |
Either NULL, or an object inheriting from class |
return. |
character string: see Value |
... |
further arguments passed to or from other methods. For |
The smoothObject
argument may be used to redraw a figure faster by recycling the predictor of the response surface returned invisibly by a previous call to filled.mapMM
.
For smoothObject=NULL
(the default), filled.mapMM
interpolates the predicted response, with sometimes unpleasant effects. For example, if one interpolates probabilities, the result may not be within [0,1], and then (say) a logarithmic Ztransf
may generate NaN values that would otherwise not occur. The smoothObject
argument may be used to overcome the default behaviour, by providing an alternative predictor.
If you have values for all predictor variables in all locations of a fine spatial grid, filled.mapMM
may not be a good choice, since it will ignore that information (see map.formula
argument). Rather, one should use predict(<fitobject>,newdata= <all predictor variables >)
to generate all predictions, and then either
spaMM.filled.contour
or some other raster functions.
The different functions are (currently) inconsistent among themselves in the way they handle the plot.title
and plot.axes
argument:
spaMM.filled.contour behaves like graphics::filled.contour
, which (1) handles arguments which are calls such as title(.)
or {axis(1);axis(2)}
; (2) ignores ...
arguments if plot.title
is missing; and (3) draws axes by default when plot.axes
is missing, given axes = TRUE
.
By contrast, filled.mapMM handles arguments which are language expressions such as produced by quote(.)
or substitute(.)
(see Examples).
mapMM can handles language expressions, but also accepts at least some calls.
filled.mapMM
by default returns invisibly the fit object predicting the interpolated response surface; however, for any non-default return.
argument (return.="raster"
would be recommended to ensure future back-compatibility), it will return a raster of values as a list with elements x
, y
and z
. map_ranef
returns invisibly a 3-column matrix containing the spatial coordinates, and the predicted effect z
on the linear predictor scale (which is also the scale of the plot, unless a Ztransf
is used). mapMM
returns invisibly a list with elements x
, y
and z
. Plots are produced as side-effects.
seaMask
for masking areas in a filled map;
https://gitlab.mbb.univ-montp2.fr/francois/spamm-ref/-/blob/master/vignettePlus/example_raster.html for more elaborate plot procedures.
data("blackcap") bfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=0.5537,phi=1.376e-05,rho=0.0544740,nu=0.6286311)) mapMM(bfit,color.palette = function(n){spaMM.colors(n,redshift=1/2)},add.map=TRUE) map_ranef(bfit) # providing argument re.form= . ~ Matern(1|longitude+latitude) if (spaMM.getOption("example_maxtime")>1) { ## filled.mapMM takes a bit longer # showing 'add.map', 'nlevels', and contour lines for 'variance' filled.mapMM(bfit, nlevels=30, add.map=TRUE, plot.axes=quote({axis(1);axis(2)}), variance="respVar", plot.title=title(main="Inferred migration propensity of blackcaps", xlab="longitude",ylab="latitude")) ## Similar plots by ggplot2: ## Not run: library(rnaturalearth) # provides sea mask through 'ne_download' function library(ggplot2) library(sp) # sea mask sea <- ne_download(scale = 10, type = 'ocean', category = "physical", returnclass = "sf") # Generation of data.frame for ggplot: rastr <- filled.mapMM(bfit, return.="raster") spdf <- data.frame(Long=rep(rastr$x, nc), Lat=rastr$y[gl(nr,nc)], z = as.vector(rastr$z)) ggplot(spdf) + geom_contour_filled(aes(Long,Lat,z=z), bins = 20) + guides(fill = "none") + geom_sf(data = sea, fill = "black") + coord_sf(ylim = range(rastr$y), xlim = range(rastr$x), expand = FALSE) ## End(Not run) } if (spaMM.getOption("example_maxtime")>3) { data("Loaloa") lfit <- fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), method="PQL", data=Loaloa, family=binomial(), fixed=list(nu=0.5,rho=2.255197,lambda=1.075)) ## longer computation requiring interpolation of 197 points ## Also illustrating effect of 'return.' argument res <- filled.mapMM(lfit,add.map=TRUE,plot.axes=quote({axis(1);axis(2)}), decorations=quote(points(pred[,coordinates],pch=15,cex=0.3)), return.="raster", # so that 'res' is a list representing a raster. plot.title=title(main="Inferred prevalence, North Cameroon", xlab="longitude",ylab="latitude")) }
data("blackcap") bfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=0.5537,phi=1.376e-05,rho=0.0544740,nu=0.6286311)) mapMM(bfit,color.palette = function(n){spaMM.colors(n,redshift=1/2)},add.map=TRUE) map_ranef(bfit) # providing argument re.form= . ~ Matern(1|longitude+latitude) if (spaMM.getOption("example_maxtime")>1) { ## filled.mapMM takes a bit longer # showing 'add.map', 'nlevels', and contour lines for 'variance' filled.mapMM(bfit, nlevels=30, add.map=TRUE, plot.axes=quote({axis(1);axis(2)}), variance="respVar", plot.title=title(main="Inferred migration propensity of blackcaps", xlab="longitude",ylab="latitude")) ## Similar plots by ggplot2: ## Not run: library(rnaturalearth) # provides sea mask through 'ne_download' function library(ggplot2) library(sp) # sea mask sea <- ne_download(scale = 10, type = 'ocean', category = "physical", returnclass = "sf") # Generation of data.frame for ggplot: rastr <- filled.mapMM(bfit, return.="raster") spdf <- data.frame(Long=rep(rastr$x, nc), Lat=rastr$y[gl(nr,nc)], z = as.vector(rastr$z)) ggplot(spdf) + geom_contour_filled(aes(Long,Lat,z=z), bins = 20) + guides(fill = "none") + geom_sf(data = sea, fill = "black") + coord_sf(ylim = range(rastr$y), xlim = range(rastr$x), expand = FALSE) ## End(Not run) } if (spaMM.getOption("example_maxtime")>3) { data("Loaloa") lfit <- fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), method="PQL", data=Loaloa, family=binomial(), fixed=list(nu=0.5,rho=2.255197,lambda=1.075)) ## longer computation requiring interpolation of 197 points ## Also illustrating effect of 'return.' argument res <- filled.mapMM(lfit,add.map=TRUE,plot.axes=quote({axis(1);axis(2)}), decorations=quote(points(pred[,coordinates],pch=15,cex=0.3)), return.="raster", # so that 'res' is a list representing a raster. plot.title=title(main="Inferred prevalence, North Cameroon", xlab="longitude",ylab="latitude")) }
mat_sqrt
is not usually directly called by users, but arguments may be passed to it through higher-level calls (see Examples).
For given matrix C, it computes a factor L such that C = L * t(L), handling issues with nearly-singular matrices. The default behavior is to try Cholesky factorization, and use eigen
if it fails.
Matrix roots are not unique (for example, they are lower triangular for t(chol(.))
, and symmetric for svd(.)
. As matrix roots are used to simulate samples under the fitted model (in particular in the parametric bootstrap implemented in fixedLRT
), this implies that for given seed of random numbers, these samples will differ with these different methods (although their distribution should be identical).
mat_sqrt(m = NULL, symSVD = NULL, try.chol = TRUE, condnum=1e12)
mat_sqrt(m = NULL, symSVD = NULL, try.chol = TRUE, condnum=1e12)
m |
The matrix whose 'root' is to be computed. This argument is ignored if |
symSVD |
A list representing the symmetric singular value decomposition of the matrix which 'root' is to be computed. Must have elements |
try.chol |
If |
condnum |
(large) numeric value. In the case |
For non-NULL m
, its matrix root, with rows and columns labelled according to the columns of the original matrix.
If eigen
was used, the symmetric singular value decomposition (a list with members u
(matrix of eigenvectors) and d
(vector of eigenvalues)) is given as attribute.
## Not run: ## try.chol argument passed to mat_sqrt ## through the '...' argument of higher-level functions ## such as HLCor, corrHLfit, fixedLRT: data("scotlip") HLCor(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), ranPars=list(rho=0.174),adjMatrix=Nmatrix,family=poisson(), data=scotlip,try.chol=FALSE) ## End(Not run)
## Not run: ## try.chol argument passed to mat_sqrt ## through the '...' argument of higher-level functions ## such as HLCor, corrHLfit, fixedLRT: data("scotlip") HLCor(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), ranPars=list(rho=0.174),adjMatrix=Nmatrix,family=poisson(), data=scotlip,try.chol=FALSE) ## End(Not run)
The Matérn correlation function describes realizations of Gaussian spatial processes with different smoothnesses
(i.e. either smooth or rugged surfaces, controlled by the parameter). It also includes a
scaling parameter and an optional 'nugget' parameter. A random effect specified in a model formula as
Matern(1|
<...>)
has pairwise correlations given by the Matérn function at the scaled Euclidean distance between coordinates specified in <...>, using “+” as separator (e.g., Matern(1|longitude+latitude)
). The Matern family can be used in Euclidean spaces of any dimension; and also for correlations on a sphere (with maximum smoothness nu=0.5
).
A syntax of the form Matern(1|longitude+latitude %in% grp)
can be used to specify a Matern random effect with independent realizations (but identical correlation parameters) for each level of the grouping variable grp
. Alternatively, the Matern(<T/F factor>|longitude+latitude)
may be used to specify Matern effects specific to individuals identified by the <T/F factor>
(see Example with females and males). In that case distinct correlation parameters are fitted for each such Matern term.
When group-specific autocorrelated random effects are fitted, it may be wise to allow for different means for each group in the Intercept (a message will point this out if the fit results for Matern or Cauchy terms suggest so).
By default, fitme
and corrHLfit
performs optimization over the and
parameters. It is possible to estimate different scaling parameters for the different Euclidean dimensions: see examples in
make_scaled_dist
.
The MaternCorr
function may be used to vizualise these correlations, using distances as input.
## Default S3 method: MaternCorr(d, rho = 1, smoothness, nu = smoothness, Nugget = NULL) # Matern(1|...)
## Default S3 method: MaternCorr(d, rho = 1, smoothness, nu = smoothness, Nugget = NULL) # Matern(1|...)
d |
A distance or a distance matrix. |
rho |
A scaling factor for distance. The 'range' considered in some formulations is the reciprocal of this scaling factor |
smoothness |
The smoothness parameter, >0. |
nu |
Same as smoothness |
Nugget |
(Following the jargon of Kriging) a parameter describing a discontinuous decrease in
correlation at zero distance. Correlation will always be 1 at |
... |
Names of coordinates, using “+” as separator (e.g., |
The correlation at distance is
where
is the
besselK
function of order .
By default the Nugget is set to 0. See one of the examples on data set Loaloa
for a fit including the estimation of the Nugget.
Scalar/vector/matrix depending on input.
Stein, M.L. (1999) Statistical Interpolation of Spatial Data: Some Theory for Kriging. Springer, New York.
See corMatern
for an implementation of this correlation function as a corSpatial
object for use with lme
or glmmPQL
.
## See examples in help("HLCor"), help("Loaloa"), help("make_scaled_dist"), etc. ## Matern correlations in 4-dimensional space: set.seed(123) randpts <- matrix(rnorm(20),nrow=5) distMatrix <- as.matrix(proxy::dist(randpts)) MaternCorr(distMatrix,nu=2) ## Group-specific random effects: if (spaMM.getOption("example_maxtime")>1.6) { data(Leuca) subLeuca <- Leuca[c(1:10,79:88),] # subset of 10 females and 10 males, for faster examples. # Independent Matern random effect with different covariance parameters for each sex: fitme(fec_div ~ sex + Matern(female|x + y) + Matern(male|x + y), data = subLeuca) # Independent Matern random effect with the same covariance parameters for each sex: fitme(fec_div ~ sex + Matern(1|x+y %in% sex),data=subLeuca) # Matern models with random-effects distinct but correlated across sexes # can also be fitted: see Matern examples in help("composite-ranef"). }
## See examples in help("HLCor"), help("Loaloa"), help("make_scaled_dist"), etc. ## Matern correlations in 4-dimensional space: set.seed(123) randpts <- matrix(rnorm(20),nrow=5) distMatrix <- as.matrix(proxy::dist(randpts)) MaternCorr(distMatrix,nu=2) ## Group-specific random effects: if (spaMM.getOption("example_maxtime")>1.6) { data(Leuca) subLeuca <- Leuca[c(1:10,79:88),] # subset of 10 females and 10 males, for faster examples. # Independent Matern random effect with different covariance parameters for each sex: fitme(fec_div ~ sex + Matern(female|x + y) + Matern(male|x + y), data = subLeuca) # Independent Matern random effect with the same covariance parameters for each sex: fitme(fec_div ~ sex + Matern(1|x+y %in% sex),data=subLeuca) # Matern models with random-effects distinct but correlated across sexes # can also be fitted: see Matern examples in help("composite-ranef"). }
Reimplements the IMRF
correlation model approximating a Matern correlation function, through a corrFamily
constructor. This allows the efficient joint estimation of the alpha
parameter of the approximating Markov random field (in principle related to the smoothness parameter of the Matern
correlation function) together with its kappa
parameter. By contrast, random effects terms specified as IMRF(1| . , model = <INLA::inla.spde2.matern result>)
assume a fixed alpha
.
Using this feature requires that the not-on-CRAN package INLA (https://www.r-inla.org) is installed so that INLA::inla.spde2.matern
can be called for each alpha
value.
# corrFamily constructor: MaternIMRFa(mesh, tpar = c(alpha = 1.25, kappa = 0.1), fixed = NULL, norm=FALSE)
# corrFamily constructor: MaternIMRFa(mesh, tpar = c(alpha = 1.25, kappa = 0.1), fixed = NULL, norm=FALSE)
mesh |
An |
and consistently with the general format of corrFamily
constructors:
tpar |
Named numeric vector: template values of the parameters of the model. Better not modified unless you know what you are doing. |
fixed |
NULL or numeric vector, to fix the parameters of this model. |
norm |
Boolean: whether to apply a normalization so that the random effect is homoscedastic (see |
A list suitable as input in the covStruct
argument, with the following elements:
f |
function returning a precision matrix for the random effect in mesh vertices; |
tpar |
template parameter vector (see general requirements of a |
Af |
function returning a matrix that implements the prediction of random effect values in data locations by interpolation of values in |
type |
specifies that the matrix returned by |
and possibly other elements which should not be considered as stable features of the return value.
Lindgren F., Rue H., Lindström J. (2011) An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73: 423-498. doi:10.1111/j.1467-9868.2011.00777.x
## Not run: if(requireNamespace("INLA", quietly = TRUE)) { data("Loaloa") mesh <- INLA::inla.mesh.2d(loc = Loaloa[, c("longitude", "latitude")], max.edge = c(3, 20)) ### Fit with fixed alpha (fit_MaternIMRF <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + MaternIMRFa(1|longitude+latitude, mesh, fixed=c(alpha=1.05)), family=binomial(), data=Loaloa, verbose=c(TRACE=interactive())) ) # For data sets with a small number of locations (as here), fitting # the Matern model as follows is faster than fitting its MaternIMRFa approximation. # Here this appears more than twofold faster fit_Matern <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + Matern(1|longitude+latitude), fixed=list(nu=0.05), # in principle similar to alpha=0.05 data=Loaloa,family=binomial()) ### Same with variable alpha (fit_MaternIMRF <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + MaternIMRFa(1|longitude+latitude, mesh), family=binomial(), data=Loaloa, verbose=c(TRACE=interactive())) ) # Comparable Matern fit: fit_Matern <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + Matern(1|longitude+latitude), init=list(nu=0.25), lower=list(nu=0), upper=list(nu=1), data=Loaloa,family=binomial()) # Note that the fitted nu and alpha parameters do not quite match each other, # and that the IMRF likelihood does not really approximate the Matern likelihood. # Mesh design would also be seen to matter. } else print("INLA must be installed to run this example.") ## End(Not run)
## Not run: if(requireNamespace("INLA", quietly = TRUE)) { data("Loaloa") mesh <- INLA::inla.mesh.2d(loc = Loaloa[, c("longitude", "latitude")], max.edge = c(3, 20)) ### Fit with fixed alpha (fit_MaternIMRF <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + MaternIMRFa(1|longitude+latitude, mesh, fixed=c(alpha=1.05)), family=binomial(), data=Loaloa, verbose=c(TRACE=interactive())) ) # For data sets with a small number of locations (as here), fitting # the Matern model as follows is faster than fitting its MaternIMRFa approximation. # Here this appears more than twofold faster fit_Matern <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + Matern(1|longitude+latitude), fixed=list(nu=0.05), # in principle similar to alpha=0.05 data=Loaloa,family=binomial()) ### Same with variable alpha (fit_MaternIMRF <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + MaternIMRFa(1|longitude+latitude, mesh), family=binomial(), data=Loaloa, verbose=c(TRACE=interactive())) ) # Comparable Matern fit: fit_Matern <- fitme( cbind(npos,ntot-npos) ~ elev1 + elev2 + elev3 + elev4 + maxNDVI1 + seNDVI + Matern(1|longitude+latitude), init=list(nu=0.25), lower=list(nu=0), upper=list(nu=1), data=Loaloa,family=binomial()) # Note that the fitted nu and alpha parameters do not quite match each other, # and that the IMRF likelihood does not really approximate the Matern likelihood. # Mesh design would also be seen to matter. } else print("INLA must be installed to run this example.") ## End(Not run)
The method
argument of the fitting functions, with possible values "ML"
, "REML"
,"PQL"
, "PQL/L"
, and so on, controls whether restricted likelihood techniques are used to estimate residual variance and random-effect parameters, and the way likelihood functions are approximated.
By default, Laplace approximations are used, as selected by "ML"
and "REML"
methods. The Laplace approximation to (log-)marginal likelihood can be expressed in terms of the joint log-likelihood of the data and the random effects (or the h-likelihood in works of Lee and Nelder). The Laplace approximation is the joint likelihood minus half the log-determinant of the matrix of second derivatives (Hessian matrix) of the negative joint likelihood with respect to the random effects (observed information matrix). The Laplace approximation to restricted likelihood (for REML estimation) is similarly defined from the Hessian matrix with respect to random effects and fixed effects (for the adventurous, spaMM allows some non-standard specification of the fixed effects included in the definition of he Hessian).
Various additional approximations have been considered. Penalized quasi-likelihood (PQL), as originally defined for GLMMs by Breslow and Clayton (1993), uses a Laplace approximation of restricted likelihood to estimate dispersion parameters, and estimates fixed effects by maximizing the joint likelihood (h-likelihood). Although PQL has been criticized as an approximation of likelihood (and actual implementations may diverge from the original version), it has some interesting inferential properties. spaMM allows one to use an ML variant of PQL, named PQL/L.
Further approximations defined by Lee, Nelder and collaborators (e.g., Noh and Lee, 2007, for some overview) may mostly be seen as laying between PQL and the full Laplace method in terms of approximation of the likelihood, and as extending them to models with non-gaussian random effects (“HGLMs”).
In practice the ML, REML, PQL and PQL/L methods should cover most (all?) needs for GLMMs, and EQL extends the PQL concept to HGLMs. method="EQL+"
stands for the EQL method of Lee and Nelder (2001). The '+' signals that it includes the d v/ d tau correction described p. 997 of that paper, while method="EQL-"
ignores it. "PQL"
is equivalent to EQL-
for GLMMs. "PQL/L"
is PQL without the leverage corrections that characterize REML estimation of random-effect parameters.
spaMM uses the observed information matrix by default since version 4.0.0. By contrast, in Laplace approximations of likelihood described in the work of Lee & Nelder, i.e. for mixed-effect models with GLM response families, the information matrix is written in terms of the GLM weights (e.g., Lee & Nelder 2001, p.1004), and is thus effectively the expected information matrix, which differs from the observed information matrix in the case of GLM families with non-canonical link (McCullagh & Nelder 1989, p.42). Therefore, the likelihood approximation based on the expected information matrix differs from the one based on the observed information matrix in the same conditions.
For non-GLM response families (currently, the negbin1
, beta_resp
and betabin
), only observed information is available (expected information would at best be quite difficult to evaluate, with no benefits). For GLM response families, use of expected information matrix can be required at a global level by setting spaMM.options(obsInfo=FALSE)
or in a specific fit by adding "exp"
as a second specifier in the method (e.g., method=c("ML","exp")
). This can be distinctly useful (in terms of speed) for fitting models with Gamma(log)
response family. Conversely, the "obs"
specifier will enforce use of observed information matrix when the alternative is set at a global level.
The method
(or HLmethod
) argument of fitting functions also accepts values of the form "HL(<...>)"
, "ML(<...>)"
and "RE(<...>)"
, e.g. method="RE(1,1)"
, which allow one to experiment with further combinations of approximations. HL and RE are equivalent (both imply an REML correction). The first '1' means that a Laplace approximation to the likelihood is used to estimate fixed effects
(a '0' would instead mean that the h likelihood is used as the objective function). The second '1' means that a Laplace approximation to the likelihood or restricted likelihood is used to estimate dispersion parameters, this approximation including the dv/d tau term specifically discussed by Lee & Nelder 2001, p. 997 (a '0' would instead mean that these terms are ignored). It is possible to enforce the EQL approximation for estimation of dispersion parameter (i.e., Lee and Nelder's (2001) method) by adding a third index with value 0. "EQL+"
is thus "HL(0,1,0)"
, while "EQL-"
is "HL(0,0,0)"
. "PQL"
is EQL-
for GLMMs. "REML"
is "HL(1,1)"
. "ML"
is "ML(1,1)"
.
Some of these distinctions make sense for GLMs, and may help in understanding idiosyncrasies of stats::glm
for Gamma GLMs. In particular (as stated in the stats::logLik
documentation) the logLik of a Gamma GLM fit by glm
differs from the exact likelihood. An "ML(0,0,0)"
approximation of true ML provides the same log likelihood as stats::logLik
. Further, the dispersion estimate returned by summary.glm
differs from the one implied by logLik
, because summary.glm
uses Pearson residuals instead of deviance residuals. This may be confusing, and no method
in spaMM tries to reproduce simultaneously these distinct features (however, spaMM_glm
may do so). The dispersion estimate returned by an "HL(.,.,0)"
fit matches what can be computed from residual deviance and residual degrees of freedom of a glm
fit, but this is not the estimate displayed by summary.glm
. The fixed effect estimates are not affected by these tinkerings.
Breslow, NE, Clayton, DG. (1993). Approximate Inference in Generalized Linear Mixed Models. Journal of the American Statistical Association 88, 9-25.
Lee, Y., Nelder, J. A. (2001) Hierarchical generalised linear models: A synthesis of generalised linear models, random-effect models and structured dispersions. Biometrika 88, 987-1006.
McCullagh, P. and Nelder, J.A. (1989) Generalized Linear Models, 2nd edition. London: Chapman & Hall.
Noh, M., and Lee, Y. (2007). REML estimation for binary data in GLMMs, J. Multivariate Anal. 98, 896-915.
This implements the procedure described by Benjamini and Gavrilov (2009) for model-selection of fixed-effect terms based on False Discovery Rate (FDR) concepts. It uses forward selection based on penalized likelihoods. The penalization for the number of parameters is distinct from that in Akaike's Information Criterion, and variable across iterations of the algorithm (but functions from the stats
package for AIC-based model-selection are still called, so that some screen messages refer to AIC).
MSFDR(nullfit, fullfit, q = 0.05, verbose = TRUE)
MSFDR(nullfit, fullfit, q = 0.05, verbose = TRUE)
nullfit |
An ML fit to the minimal model to start the forward selection from; an object of class |
fullfit |
An ML fit to the maximal model; an object of class |
q |
Nominal error rate of the underlying FDR procedure (expected proportion of
incorrectly rejected null out of the rejected). Benjamini and Gavrilov (2009) recommend |
verbose |
Whether to print information about the progress of the procedure. |
The fit of the final selected model; an object of class HLfit
.
A simple forward selection procedure based on false discovery rate control. Ann. Appl. Stat, 3, 179-198 (2009).
if (spaMM.getOption("example_maxtime")>1.4) { data("wafers") nullfit <- fitme(y~1+(1|batch), data=wafers,family=Gamma(log)) fullfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log)) MSFDR(nullfit=nullfit,fullfit=fullfit) }
if (spaMM.getOption("example_maxtime")>1.4) { data("wafers") nullfit <- fitme(y~1+(1|batch), data=wafers,family=Gamma(log)) fullfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch), data=wafers, family=Gamma(log)) MSFDR(nullfit=nullfit,fullfit=fullfit) }
spaMM can fit random-effect terms of the forms considered by Lindgren et al. (2011) or Nychka et al. (2015, 2018). The random effects considered here all involve a multivariate Gaussian random effect over a lattice, from which the random-effect value in any spatial position is determined by interpolation of values on the lattice. IMRF stands for Interpolated Markov Random Field because the specific process considered on the lattice is currently known as a Gaussian Markov Random Field (see the Details for further information). Lindgren et al. considered irregular lattices designed to approximate of the Matern correlation model with fixed smoothness <= 2, while Nychka et al. considered regular grids.
The correlation model of Lindgren et al. (2011) can be fitted by spaMM by declaring an IMRF
random effect term in the model formula, with a model
argument in the right-hand side whose value is the result of INLA::inla.spde2.matern
(or INLA::inla.spde2.pcmatern
) for given smoothness. The spaMM functions for such a fit do not call INLA functions. Alternatively, the same model with variable smoothness can be fitted by declaring a corrFamily
term whose structure is described through the MaternIMRFa
function, whose respective documentations should be considered for more details. In the latter case INLA::inla.spde2.matern
is called internally by spaMM
. The correlation models thus defined are fitted by the same methods as other models in spaMM.
Regular lattices can also be declared as an IMRF
term (with arguments distinct from model
). The multIMRF
syntax implements the multiresolution model of Nychka et al. Any multIMRF
term in a formula is immediately converted to IMRF
terms for regular grids wih different step sizes. This has distinct implications for controlling the parameters of these or other random effects in the model by init
or fixed
values: see Details if you need such control.
# IMRF( 1 | <coordinates>, model, nd, m, no, ce, ...) # multIMRF( 1 | <coordinates>, levels, margin, coarse=10L, # norm=TRUE, centered=TRUE )
# IMRF( 1 | <coordinates>, model, nd, m, no, ce, ...) # multIMRF( 1 | <coordinates>, levels, margin, coarse=10L, # norm=TRUE, centered=TRUE )
model |
An |
levels |
integer; Number of levels in the hierarchy, i.e. number of component IMRFs. |
margin , m
|
integer; width of the margin, as a number of additional grid points on each side (applies to all levels of the hierarchy). |
coarse |
integer; number of grid points (excluding the margins) per dimension for the coarsest IMRF. The number of grids steps nearly doubles with each level of the hierarchy (see Details). |
nd |
integer; number of grid steps (excluding the margins) per dimension for the given IMRF. |
norm , no
|
Boolean; whether to apply normalization (see Details), or not. |
centered , ce
|
Boolean; whether to center the grid in all dimensions, or not. |
... |
Not documented, for programming purposes |
Formulation of the covariance models:
Gaussian Markov Random Field (MRF) and conditional autoregressive models are essentially the same thing, apart from details of specification.
adjacency
and AR1
random effects can be seen as specific MRFs.
The common idea is the Markov-like property that the distribution of each element of the random-effect b, given values of a few specific elements (the “neighbours” of
), is independent of other elements (i.e., of non-neighbours). The non-zero non-diagonal elements of a precision matrix characterize the neighbours.
Given the inferred vector b of values of the MRF on the lattice, the interpolation of the MRF in any focal point is of the form Ab where each row of A weights the values of b according to the position of the focal point relative to the vertices of the lattice. Following the original publications, *
for grids given by model=<inla.spde2 object>
, the non-zero weights are the barycentric coordinates of the focal point in the enclosing triangle from the mesh triangulation (points from outside the mesh would have zero weights, so the predicted effect Ab=0); *
for regular grids (NULL model
), the weights are computed as <Wendland function>(<scaled Euclidean distances between focal point and vertices>).
The IMRF
model defines both a lattice in space, the precision matrix for a Gaussian MRF over this lattice, and the A matrix of weights. The full specification of the MRF on irregular lattices is complex. The (
kappa
) parameter considered by spaMM
is the scale parameter considered by Lindgren et al and comparable to the
scale factor of the Matérn model. The
argument of the
INLA::inla.spde2.matern
controls the smoothness of the approximated Matern model, as ) where
is the dimension of the space.
Correlation models created by
INLA::inla.spde2.pcmatern
are handled so as to give the same correlation values as when INLA::inla.spde2.matern
is used with the same mesh
and alpha
argument (thus, the extra functionalities of “pc
”matern
are ignored).
Not all options of the INLA functions may be compatible or meaningful when used with spaMM (only the effects of alpha
and cutoff
have been checked).
Normalization:
For the MRFs on default regular grids (missing model
argument), the precision matrix is defined (up to a variance parameter) as M'M where the diagonal elements of M are 4+
and the
for the four nearest neighbours are -1 (note that M'M involves more than these four neighbours). The precision matrix defined in this way is the inverse of an heteroscedastic covariance matrix C, but (following Nychka et al.) by default a normalization is applied so that the random effect in each data position is homoscedastic (the precision matrix for the latent effect in grid positions is not modified, but it is the A matrix of weights which is is modified). As for other random effects, the variance is further controlled by a multiplicative factor
.
Without normalization, the covariance matrix of the random effect in data locations is ALL'A' (A being the above-described weight matrix, and L is a “square root” of C), and AL is the original “design matrix” of the random effect.
may then be quite different from the marginal variance of the random effect, and is difficult to describe in a simple way.
For normalization, A is modified as WA where W is a diagonal matrix such that WAL is a correlation matrix (WALL'A'W' has unit diagonal); then,
is the marginal variance of the random effect.
For irregular grids specified using the model
argument, the precision matrix described by this object is also the inverse of an heteroscedastic covariance matrix, but here (again following original publicatiosn such as Lindgren at al. 2011) the normalization is not applied by default (and was not even an option before version 4.3.23). But for ease of presentation and interpretation, if for no other reason, the normalized model may be preferable.
Details for rectangular grids:
By default (meaning in particular that model
is not used to specify a lattice defined by the INLA procedures), the IMRF lattice is rectangular (currently the only option) and is made of a core lattice, to which margins of margin
steps are added on each side. The core lattice is defined as follows: in each of the two spatial dimensions, the range of axial coordinates is determined. The largest range is divided in nd-1
steps, determining nd
values and step length . The other range is divided in steps of the same length
. If it extends over (say)
, a grid of 2 steps and 3 values is defined, and by default centered on the range (the extreme points therefore typically extend slightly beyond the grid, within the first of the additional steps defined by the
margin
; if not centered, the grid start from the lower coordinate of the range).
multIMRF
implements multilevel IMRFs. It defines a sequence of IMRFs, with progressively finer lattices, a common value
hy_kap
for these IMRFs, and a single variance parameter hy_lam
that determines values decreasing by a factor of 4 for successive IMRF terms. By default, each component
IMRF
is normalized independently as described above (as in Nychka et al. 2019), and hy_lam
is the sum of the variances of these terms (e.g., if there are three levels and hy_lam=1
, the successive variances are (1,1/4,1/16)/(21/16) ). The nd
of the first IMRF is set to the coarse
value, and its lattice is defined accordingly. If coarse=4
and margin=5
, a grid of 14 coordinates is therefore defined over the largest range. In the second IMRF, the grid spacing is halved, so that new steps are defined halfway between the previous ones (yielding a grid of 27 step in the widest range). The third IMRF proceeds from the second in the same way, and so on.
To control initial or fixed values of multIMRF
and variance parameters, which are hyper-parameter controlling several
IMRF
terms, the hyper
syntax shown in the Examples should be used. hyper
is a nested list whose possible elements are named "1"
, "2"
, ... referring to successive multIMRF
terms in the input formula, not to successive random effect in the expanded formula with distinct IMRF terms (see Examples). But the different IMRF terms should be counted as distinct random effects when controlling other parameters (e.g., for fixing the variances of other random effects).
D. Nychka, S. Bandyopadhyay, D. Hammerling, F. Lindgren, S. Sain (2015) A multiresolution gaussian process model for the analysis of large spatial datasets. Journal of Computational and Graphical Statistics 24 (2), 579-599. doi:10.1080/10618600.2014.914946
D. Nychka, D. Hammerling, Mitchel. Krock, A. Wiens (2018) Modeling and emulation of nonstationary Gaussian fields. Spat. Stat. 28: 21-38. doi:10.1016/j.spasta.2018.08.006
Lindgren F., Rue H., Lindström J. (2011) An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach Journal of the Royal Statistical Society: Series B (Statistical Methodology), 73: 423-498. doi:10.1111/j.1467-9868.2011.00777.x
if (spaMM.getOption("example_maxtime")>6) { data("blackcap") ## toy examples; but IMRF may be useful only for much larger datasets ## and when using the 'cutoff' parameter of INLA::inla.mesh.2d() ########################## Irregular lattice specified by 'model': # data("small_spde") ## load object of class 'inla.spde2', created and saved by : # spd <- sp::SpatialPointsDataFrame(coords = blackcap[, c("longitude", "latitude")], # data = blackcap) # small_mesh <- INLA::inla.mesh.2d(loc = INLA::inla.mesh.map(sp::coordinates(spd)), # max.n=100, # only for demonstration purposes # max.edge = c(3, 20)) # small_spde <- INLA::inla.spde2.matern(small_mesh) # save(small_spde, file="small_spde.RData", version=2) # fit_SPDE <- fitme(migStatus ~ means + IMRF(1|longitude+latitude, model=small_spde), data=blackcap) ########################## Regular lattices: # #Using 'hyper' to control fixed hyper-parameters # (mrf <- fitme(migStatus ~ 1 + (1|pos) + multIMRF(1|longitude+latitude,margin=5,levels=2), data=blackcap, fixed=list(phi=1,lambda=c("1"=0.5), hyper=list("1"=list(hy_kap=0.1,hy_lam=1)))) ) # Using 'hyper' to control initial hyper-parameters # (mrf <- fitme(migStatus ~ 1 + multIMRF(1|longitude+latitude,margin=5,levels=2), data=blackcap, method="ML", fixed =list(phi=1), init=list(hyper=list("1"=list(hy_kap=0.1,hy_lam=1)))) ) # *Independent* IMRF terms with default rectangular lattice (often giving dubious results) # (mrf <- fitme(migStatus ~ 1 + IMRF(1|longitude+latitude,margin=5, nd=4L) + IMRF(1|longitude+latitude,margin=5, nd=7L), data=blackcap, fixed=list(phi=1,lambda=c(1/4,1/16), corrPars=list("1"=list(kappa=0.1),"2"=list(kappa=0.1))))) }
if (spaMM.getOption("example_maxtime")>6) { data("blackcap") ## toy examples; but IMRF may be useful only for much larger datasets ## and when using the 'cutoff' parameter of INLA::inla.mesh.2d() ########################## Irregular lattice specified by 'model': # data("small_spde") ## load object of class 'inla.spde2', created and saved by : # spd <- sp::SpatialPointsDataFrame(coords = blackcap[, c("longitude", "latitude")], # data = blackcap) # small_mesh <- INLA::inla.mesh.2d(loc = INLA::inla.mesh.map(sp::coordinates(spd)), # max.n=100, # only for demonstration purposes # max.edge = c(3, 20)) # small_spde <- INLA::inla.spde2.matern(small_mesh) # save(small_spde, file="small_spde.RData", version=2) # fit_SPDE <- fitme(migStatus ~ means + IMRF(1|longitude+latitude, model=small_spde), data=blackcap) ########################## Regular lattices: # #Using 'hyper' to control fixed hyper-parameters # (mrf <- fitme(migStatus ~ 1 + (1|pos) + multIMRF(1|longitude+latitude,margin=5,levels=2), data=blackcap, fixed=list(phi=1,lambda=c("1"=0.5), hyper=list("1"=list(hy_kap=0.1,hy_lam=1)))) ) # Using 'hyper' to control initial hyper-parameters # (mrf <- fitme(migStatus ~ 1 + multIMRF(1|longitude+latitude,margin=5,levels=2), data=blackcap, method="ML", fixed =list(phi=1), init=list(hyper=list("1"=list(hy_kap=0.1,hy_lam=1)))) ) # *Independent* IMRF terms with default rectangular lattice (often giving dubious results) # (mrf <- fitme(migStatus ~ 1 + IMRF(1|longitude+latitude,margin=5, nd=4L) + IMRF(1|longitude+latitude,margin=5, nd=7L), data=blackcap, fixed=list(phi=1,lambda=c(1/4,1/16), corrPars=list("1"=list(kappa=0.1),"2"=list(kappa=0.1))))) }
These functions facilitate the conversion and analysis of multinomial data as as series of nested binomial data.
The main interface is the multi
“family”, to be used in the family
argument of the fitting functions.
Fits using it call binomialize
, which can be called directly to check how the data are converted to nested binomial data, and to use these data directly.
The fitted.HLfitlist
method of the fitted
generic function returns a matrix of fitted multinomial probabilities.
The logLik.HLfitlist
method of the logLik
generic function returns a log-likelihood for the joint fits.
multi(binResponse=c("npos","nneg"),binfamily=binomial(),input="types",...) binomialize(data,responses,sortedTypes=NULL,binResponse=c("npos","nneg"), depth=Inf,input="types") ## S3 method for class 'HLfitlist' fitted(object, version=2L, ...) ## S3 method for class 'HLfitlist' logLik(object,which,...)
multi(binResponse=c("npos","nneg"),binfamily=binomial(),input="types",...) binomialize(data,responses,sortedTypes=NULL,binResponse=c("npos","nneg"), depth=Inf,input="types") ## S3 method for class 'HLfitlist' fitted(object, version=2L, ...) ## S3 method for class 'HLfitlist' logLik(object,which,...)
data |
The data frame to be analyzed. |
object |
A list of binomial fits returned by a multinomial analysis |
responses |
column names of the data, such that |
sortedTypes |
Names of multinomial types, i.e. levels of the multinomial response factors. Their order determines which types are taken first to define the nested binomial samples. By default, the most common types are considered first. |
binResponse |
The names to be given to the number of “success” and “failures” in the binomial response. |
depth |
The maximum number of nested binomial responses to be generated from the multinomial data. |
binfamily |
The family applied to each binomial response. |
input |
If |
which |
Which element of the |
version |
Integer, for |
... |
Other arguments passed from or to other functions. |
A multinomial response, say counts 17, 13, 25, 8, 3, 1 for types type1
to type6
, can be represented as a series of nested binomials
e.g. type1
against others (17 vs 50) then among these 50 others, type2
versus others (13 vs 37), etc.
The binomialize
function generates such a representation. By default the representation considers types in decreasing order of the number of positives, i.e. first type3
against others (25 vs 42), then type1
against others within these 42, etc. It stops if it has reached depth
nested binomial responses. This can be modified by the sortedTypes
argument, e.g. sortedTypes=c("type6","type4","type2")
.
binomialize
returns a list of data frames which can be directly provided as a data
argument for the fitting functions, with binomial response.
Alternatively, one can provide the multinomial response data frame, which will be internally converted to nested binomial data if the family
argument is a call to multinomial
(see Examples).
For mixed models, the multinomial data can be fitted to a model with the same correlation parameters, and either the same or different variances of random effects, for all binomial responses. Which analysis is performed depends on whether the variances are fitted by “outer optimization” or by HLfit
's “inner iterative” algorithm, as controlled by the init
or init.corrHLfit
arguments (see Examples). These initial values therefore affect the definition of the model being fitted.
corrHLfit
will fit different variances by default. Adding an init.corrHLfit
will force estimation of a single variance across models. fitme
's default optimization strategy is more complex, and has changed and still change over versions. This creates a back-compatibility issue where the model to be fitted may change over versions of spaMM. To avoid that, it is strongly advised to use an explicit initial value when fitting a multi
model by fitme
.
binomialize
returns a list of data frames appropriate for analysis as binomial response. Each data frame contains the original one plus
two columns named according to binResponse
.
The main fitting functions, when called on a model with family=multi(.)
, return an object of class HLfitlist
, which is a list with attributes. The list elements are fits of the nested binomial models (objects of class HLfit
). The attributes provide additional information about the overall multinomial model, such as global log-likelihood values and other information properly extracted by the how()
function.
multi
is a function that returns a list, but users may never need to manipulate this output.
fitted.HLfitlist
returns a matrix. The current default version=2L
provides meaningful fitted values (predicted multinomial frequencies for each response type) even for data rows where the nested binomial fit for a type had no response information remaining. By contrast, the first version provided a matrix with 0
s for these row*fit combinations, except for the last column; in many cases this may be confusing.
## Adding colour to the famous 'iris' dataset: iriscol <- iris set.seed(123) # Simulate colours, then fit colour frequencies: iriscol$col <- sample(c("yellow", "purple", "blue"),replace = TRUE, size = nrow(iriscol), prob=c(0.5,0.3,0.2)) colfit <- fitme(cbind(npos,nneg) ~ 1+(1|Species), family=multi(responses="col"), data=iriscol, init=list(lambda=NA)) # note warning if no 'init'... head(fitted(colfit)) # To only generate the binomial datasets: binomialize(iriscol,responses="col") ## An example considering pseudo-data at one diploid locus for 50 individuals set.seed(123) genecopy1 <- sample(4,size=50,prob=c(1/2,1/4,1/8,1/8),replace=TRUE) genecopy2 <- sample(4,size=50,prob=c(1/2,1/4,1/8,1/8),replace=TRUE) alleles <- c("122","124","126","128") genotypes <- data.frame(type1=alleles[genecopy1],type2=alleles[genecopy2]) ## Columns "type1","type2" each contains an allele type => input is "types" (the default) datalist <- binomialize(genotypes,responses=c("type1","type2")) ## two equivalent fits: f1 <- HLfit(cbind(npos,nneg)~1,data=datalist, family=binomial()) f2 <- HLfit(cbind(npos,nneg)~1,data=genotypes, family=multi(responses=c("type1","type2"))) fitted(f2) if (spaMM.getOption("example_maxtime")>1.7) { ##### Control of lambda estimation over different binomial submodels genoInSpace <- data.frame(type1=alleles[genecopy1],type2=alleles[genecopy2], x=runif(50),y=runif(50)) method <- "PQL" # for faster exampple ## Fitting distinct variances for all binomial responses: multifit <- corrHLfit(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), ranFix=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 3 multifit <- fitme(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init=list(lambda=NaN), # forcing 'inner' estimation for fitme fixed=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 3 ## Fitting the same variance for all binomial responses: multifit <- fitme(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init=list(lambda=NA), # forcing 'outer' estimation for fitme fixed=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 1 multifit <- corrHLfit(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init.corrHLfit=list(lambda=1), # forcing 'outer' estimation for corrHLfit ranFix=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 1 }
## Adding colour to the famous 'iris' dataset: iriscol <- iris set.seed(123) # Simulate colours, then fit colour frequencies: iriscol$col <- sample(c("yellow", "purple", "blue"),replace = TRUE, size = nrow(iriscol), prob=c(0.5,0.3,0.2)) colfit <- fitme(cbind(npos,nneg) ~ 1+(1|Species), family=multi(responses="col"), data=iriscol, init=list(lambda=NA)) # note warning if no 'init'... head(fitted(colfit)) # To only generate the binomial datasets: binomialize(iriscol,responses="col") ## An example considering pseudo-data at one diploid locus for 50 individuals set.seed(123) genecopy1 <- sample(4,size=50,prob=c(1/2,1/4,1/8,1/8),replace=TRUE) genecopy2 <- sample(4,size=50,prob=c(1/2,1/4,1/8,1/8),replace=TRUE) alleles <- c("122","124","126","128") genotypes <- data.frame(type1=alleles[genecopy1],type2=alleles[genecopy2]) ## Columns "type1","type2" each contains an allele type => input is "types" (the default) datalist <- binomialize(genotypes,responses=c("type1","type2")) ## two equivalent fits: f1 <- HLfit(cbind(npos,nneg)~1,data=datalist, family=binomial()) f2 <- HLfit(cbind(npos,nneg)~1,data=genotypes, family=multi(responses=c("type1","type2"))) fitted(f2) if (spaMM.getOption("example_maxtime")>1.7) { ##### Control of lambda estimation over different binomial submodels genoInSpace <- data.frame(type1=alleles[genecopy1],type2=alleles[genecopy2], x=runif(50),y=runif(50)) method <- "PQL" # for faster exampple ## Fitting distinct variances for all binomial responses: multifit <- corrHLfit(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), ranFix=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 3 multifit <- fitme(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init=list(lambda=NaN), # forcing 'inner' estimation for fitme fixed=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 3 ## Fitting the same variance for all binomial responses: multifit <- fitme(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init=list(lambda=NA), # forcing 'outer' estimation for fitme fixed=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 1 multifit <- corrHLfit(cbind(npos,nneg)~1+Matern(1|x+y),data=genoInSpace, family=multi(responses=c("type1","type2")), init.corrHLfit=list(lambda=1), # forcing 'outer' estimation for corrHLfit ranFix=list(rho=1,nu=0.5), method=method) length(unique(unlist(lapply(multifit, get_ranPars, which="lambda")))) # 1 }
Motivation: In a multivariate-response model fitted by fitmv
, one may wish to fit a random-coefficient term appearing in s submodels, that is a random effect with realized values for each of these submodels and each group, with values possibly correlated among submodels within groups. Hence one might wish to specify it as a term of the form (<submodel>|group)
, where <submodel>
would represent a factor for the s submodels. But the data are not expected to contain a factor for these submodels, so such a syntax would not work without substantial data reshaping. Instead, this effect can be stated as mv(...)
where the ...
are the indices of the submodels here the random effect appears. For example if submodels 2 and 3 include this random-coefficient term, the term can be specified as (mv(2,3)|group)
.
Composite random effects, as defined in spaMM-package
and illustrated by a bivariate-response quantitative-genetic model (Examples in Gryphon
), combine correlations among response variables of different submodels as specified by mv()
, and correlations within each response variable(as specified by a relatedness matrix in the same example).
The mv(...)
expression is treated as a factor for all purposes, meaning or example that (0+mv(2,3)|group)
can also be used, leading (as for any factor) to an alternative parametrization of the same random-coefficient model (see Examples). The parametrization through (0+mv...)
is generally recommended as its results are easier to interpret. The random-effect term is treated as a random-coefficient term for all purposes, meaning for example that fixed values can be specified for its parameters using the ranCoefs
syntax (see Examples).
# mv(...)
# mv(...)
... |
Indices of all the submodels (possibly more than two) where the random effect involving this virtual factor appears. |
Not a function, hence no return value. In the summary of the fit, levels for the different submodels s within each group are labelled .mv
s.
Gryphon
for example of composite random effects.
The X2X
argument of fitmv
for fixed effects shared across sub-models.
if (spaMM.getOption("example_maxtime")>1.1) { ## data preparation data("wafers") me <- fitme(y ~ 1+(1|batch), family=Gamma(log), data=wafers, fixed=list(lambda=0.2)) set.seed(123) wafers$y1 <- simulate(me, type="marginal") wafers$y2 <- simulate(me, type="marginal") ## fits (fitmv1 <- fitmv( submodels=list(mod1=list(formula=y1~X1+(mv(1,2)|batch), family=Gamma(log)), mod2=list(formula=y2~X1+(mv(1,2)|batch), family=Gamma(log))), data=wafers)) # alternative '0+' parametrization of the same model: (fitmv2 <- fitmv( submodels=list(mod1=list(formula=y1~X1+(0+mv(1,2)|batch), family=Gamma(log)), mod2=list(formula=y2~X1+(0+mv(1,2)|batch), family=Gamma(log))), data=wafers)) # relationship between the *correlated* effects of the two fits ranef(fitmv2)[[1]][,2]-rowSums(ranef(fitmv1)[[1]]) # ~ 0 # fit with given correlation parameter: update(fitmv2, fixed=list(ranCoefs=list("1"=c(NA,-0.5,NA)))) }
if (spaMM.getOption("example_maxtime")>1.1) { ## data preparation data("wafers") me <- fitme(y ~ 1+(1|batch), family=Gamma(log), data=wafers, fixed=list(lambda=0.2)) set.seed(123) wafers$y1 <- simulate(me, type="marginal") wafers$y2 <- simulate(me, type="marginal") ## fits (fitmv1 <- fitmv( submodels=list(mod1=list(formula=y1~X1+(mv(1,2)|batch), family=Gamma(log)), mod2=list(formula=y2~X1+(mv(1,2)|batch), family=Gamma(log))), data=wafers)) # alternative '0+' parametrization of the same model: (fitmv2 <- fitmv( submodels=list(mod1=list(formula=y1~X1+(0+mv(1,2)|batch), family=Gamma(log)), mod2=list(formula=y2~X1+(0+mv(1,2)|batch), family=Gamma(log))), data=wafers)) # relationship between the *correlated* effects of the two fits ranef(fitmv2)[[1]][,2]-rowSums(ranef(fitmv1)[[1]]) # ~ 0 # fit with given correlation parameter: update(fitmv2, fixed=list(ranCoefs=list("1"=c(NA,-0.5,NA)))) }
Returns a GLM family
object for negative-binomial model with variance quadratically related to the mean : variance=
/shape, where the shape parameter need or need not be specified, depending on usage. The zero-truncated variant can be specified as
negbin2(., trunc = 0L)
. See negbin1
for the alternative negative-binomial model with variance “linearly” related to the mean.
A fixed-effect residual-dispersion model can be fitted, using the resid.model
argument, which is used to specify the form of the logarithm of the shape parameter. Thus the variance of the response become /
exp(<specified linear expression>)
.
negbin(.)
is an alias for negbin2(.)
(truncated or not), and Tnegbin(.)
is an alias for negbin2(., trunc = 0L)
.
# (the shape parameter is actually not requested unless this is used in a glm() call) # negbin2(shape = stop("negbin2's 'shape' must be specified"), link = "log", trunc = -1L, LLgeneric = TRUE) # For use with glm(), both negbin2's 'shape' and glm's method="llm.fit" are needed. # alias with distinct arguments: Tnegbin(shape = stop("Tnegbin's 'shape' must be specified"), link = "log")
# (the shape parameter is actually not requested unless this is used in a glm() call) # negbin2(shape = stop("negbin2's 'shape' must be specified"), link = "log", trunc = -1L, LLgeneric = TRUE) # For use with glm(), both negbin2's 'shape' and glm's method="llm.fit" are needed. # alias with distinct arguments: Tnegbin(shape = stop("Tnegbin's 'shape' must be specified"), link = "log")
shape |
Shape parameter of the underlying Gamma distribution: the present negative binomial distribution can be represented as a Poisson-Gamma mixture, where the conditional Poisson mean is |
link |
log, sqrt or identity link, specified by any of the available ways for GLM links (name, character string, one-element character vector, or object of class |
trunc |
Either |
LLgeneric |
For development purposes, not documented. |
shape
is the parameter of McCullagh and Nelder (1989, p.373) and the
theta
parameter of Venables and Ripley (2002, section 7.4). The latent Gamma variable has mean 1 and variance 1/shape.
The name NB_shape
should be used to set values of shape in optimization control arguments of the fitting functions (e.g., fitme(.,init=list(NB_shape=1))
); but fixed values are set by the shape
argument.
The returned family object is formally suitable for usage with glm
if the shape
argument is specified, but such usage is not recommended as it will lead to incorrect results for the zero-truncated case.
A family object with structure similar to stats::
family object but with additional member functions for usage with spaMM fitting functions.
McCullagh, P. and Nelder, J.A. (1989) Generalized Linear Models, 2nd edition. London: Chapman & Hall.
Venables, W. N. and Ripley, B. D. (2002) Modern Applied Statistics with S-PLUS. Fourth Edition. Springer.
## Fitting negative binomial model with estimated scale parameter: data("scotlip") fitme(cases~I(prop.ag/10)+offset(log(expec)),family=negbin(), data=scotlip) negfit <- fitme(I(1+cases)~I(prop.ag/10)+offset(log(expec)),family=Tnegbin(), data=scotlip) simulate(negfit,nsim=3)
## Fitting negative binomial model with estimated scale parameter: data("scotlip") fitme(cases~I(prop.ag/10)+offset(log(expec)),family=negbin(), data=scotlip) negfit <- fitme(I(1+cases)~I(prop.ag/10)+offset(log(expec)),family=Tnegbin(), data=scotlip) simulate(negfit,nsim=3)
Returns a family
object suitable as a fitme
argument for fitting negative-binomial models with variance linearly (affinely) related to the mean : variance=
/shape, where the shape parameter need or need not be specified, depending on usage.
The model described by such a family is characterized by a linear predictor, a link function, and such a negative-binomial model for the residual variation. The zero-truncated variant of this family is also handled.
A fixed-effect residual-dispersion model can be fitted, using the resid.model
argument, which is used to specify the form of the logarithm of the shape parameter. Thus the variance of the response become /
exp(<specified linear expression>)
.
negbin1(shape = stop("negbin1's 'shape' must be specified"), link = "log", trunc = -1L)
negbin1(shape = stop("negbin1's 'shape' must be specified"), link = "log", trunc = -1L)
shape |
Parameter controlling the mean-variance relationship of the |
link |
log, sqrt or identity link, specified by any of the available ways for GLM links (name, character string, one-element character vector, or object of class |
trunc |
Either |
The name NB_shape
should be used to set values of shape in optimization control arguments of the fitting functions (e.g., fitme(.,init=list(NB_shape=1))
); but fixed values are set by the shape
argument.
The family should not be used as a glm
argument as the results would not be correct.
A list, formally of class c("LLF", "family")
. See LL-family
for details about the structure and usage of such objects.
Examples in LL-family
. resid.model
for an example with a residual-dispersion model.
Computes by numerical derivation the observed information matrix for (ideally) all parameters for mean response model, that is, the matrix of second derivatives of negative log likelihood.
The default value of the which
argument shows all classes of parameters that should be handled, including random-effect parameters (lambda
, ranCoefs
, corrPars
, and hyper
), residual dispersion parameters (phi
, NB_shape
for negbin1
and negbin2
, and beta_prec
for beta_resp
and betabin
), and fixed-effect coefficients (beta
).
Model fits including a phi-resid.model
are not fully handled, in two ways: the information matrix does not include their parameters; and if the residual dispersion model include random effects, there is good reason for the numInfo
calculation to detect that the fit has not maximized marginal likelihood with respect to most parameters.
numInfo(fitobject, transf = FALSE, which = NULL, check_deriv = TRUE, sing=1e-05, verbose=FALSE, refit_hacks=list(), ...)
numInfo(fitobject, transf = FALSE, which = NULL, check_deriv = TRUE, sing=1e-05, verbose=FALSE, refit_hacks=list(), ...)
fitobject |
Fit object returned by a spaMM fitting function. |
transf |
Whether to perform internal computations on a transformed scale (but computation on transformed scale may be implemented for fewer classes of models than default computation). |
which |
NULL, or character vector giving the sets of parameters with respect to which derivatives are to be computed.
The NULL default is equivalent to |
check_deriv |
Boolean; whether to perform some checks for possible problems (see Details). |
sing |
numeric value, or |
verbose |
Boolean: whether to print (as a list) the estimates of the parameters for which the Hessian will be computed, additional information about possibly ignored parameters, possible misuse of REML fits, and a (sort of) progress bar if the procedure is expected to last more than a few seconds. |
refit_hacks |
list of arguments; its name anticipates that it might allow hazardous manipulations in a later version of spaMM. But currently only the innocuous element |
... |
Arguments passed to |
The computation of a second derivatives is not necessarily meaningful if a first derivative does not vanish at the parameter estimates. This may occur in particular when the objective function (say, marginal likelihood) is maximized at a boundary of the parameter space (say, at zero for lambda
estimates). Further, for such values at a boundary, only one-sided derivatives can be computed, and this is not handled by numDeriv::hessian
. So, some checks may be requested to detect non-zero gradients and parameters estimated at their boundaries. The boundary checks are currently performed for lambda
and ranCoefs
estimates, if check_deriv
is set to TRUE or to NULL. Other parameters are not (yet) checked, so numInfo
may sometimes fails when such other parameter estimates are at a boundary. If check_deriv
is set to TRUE, an additional systematic check of the gradient with respect to all target parameters is performed.
NULL or a matrix.
NULL is returned if no parameter is found with respect to which a numerical information “should” be computed (where what should be done depends on the which
and check_derivs
arguments).
Otherwise a matrix is returned, with an eigvals
attribute if sing
was non-zero. This attribute is a numeric vector of eigenvalues of the matrix. If some eigenvalue(s) were lower than sing
, the vector has additional class "singeigs"
so that its printing is controlled by an ad-hoc print.singeigs
method highlighting the small eigenvalues.
data("wafers") lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) numinfo <- numInfo(lmmfit) (SEs <- sqrt(diag(solve(numinfo)))) # # => beta SEs here equal to conditional SEs shown by fit summary. # Other SEs can be compared to the approximate ones # for log(phi) and log(lambda), given by # # update(lmmfit, control=list(refit=TRUE)) # # => 1118*0.5289 and 10840*0.1024 data("blackcap") maternfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) numInfo(maternfit)
data("wafers") lmmfit <- fitme(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),data=wafers) numinfo <- numInfo(lmmfit) (SEs <- sqrt(diag(solve(numinfo)))) # # => beta SEs here equal to conditional SEs shown by fit summary. # Other SEs can be compared to the approximate ones # for log(phi) and log(lambda), given by # # update(lmmfit, control=list(refit=TRUE)) # # => 1118*0.5289 and 10840*0.1024 data("blackcap") maternfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap) numInfo(maternfit)
Allow the user to set and examine a variety of options which affect operations of the spaMM package.
spaMM.options(..., warn = TRUE) spaMM.getOption(x)
spaMM.options(..., warn = TRUE) spaMM.getOption(x)
x |
a character string holding an option name. |
warn |
Boolean: whether to warn if a previously undefined options is being defined (a protection against typos). |
... |
A named value or a list of named values. The following values, with their defaults,
are used in
and many other undocumented values for programming or development purposes. Additional options without default values can also be used (e.g., see |
spaMM.options() provides an interface for changing maximal values of parameters of the Matérn correlation function. However, it is not recommended to change these values unless a spaMM message specifically suggests so.
By default spaMM use Iteratively Reweighted Least Squares (IRLS) methods to estimate fixed-effect parameters (jointly with predictions of random effects). However, a Levenberg-Marquardt algorithm, as described by Nocedal & Wright (1999, p. 266), is also implemented. The Levenberg-Marquardt algorithm is designed to optimize a single objective function with respect to all its parameters. It is thus well suited to compute a PQL fit, which is based on maximization of a single function, the h-likelihood. By contrast, in a fit of a mixed model by (RE)ML, one computes jointly fixed-effect estimates that maximizes marginal likelihood, and random-effect values that maximize h-likelihood given the fixed-effect estimates. The gradient of marginal likelihood with respect to fixed-effect coefficients does not generally vanishes at the solution (although it remains close to zero except in “difficult” cases with typically little information in the data). The Levenberg-Marquardt algorithm is not directly applicable in this case, as it may produce random-effect values that increases marginal likelihood rather than h-likelihood. The (RE)ML variant of the algorithm implemented in spaMM may therefore use additional nested h-likelihood-maximizing steps for correcting random-effect values. In version 3.1.0 this variant was revised for improved performance in difficult cases.
For spaMM.getOption
, the current value set for option x
, or
NULL
if the option is unset.
For spaMM.options()
, a list of all set options. For
spaMM.options(<name>)
, a list of length one containing the set value,
or NULL
if it is unset. For uses setting one or more options,
a list with the previous values of the options changed (returned
invisibly).
Jorge Nocedal and Stephen J. Wright (1999) Numerical Optimization. Springer-Verlag, New York.
spaMM.options() spaMM.getOption("example_maxtime") ## Not run: spaMM.options(maxLambda=1e06) ## End(Not run)
spaMM.options() spaMM.getOption("example_maxtime") ## Not run: spaMM.options(maxLambda=1e06) ## End(Not run)
This example illustrates how to use spaMM for quantitative genetic analyses. spaMM appears competitive in terms of speed for GLMMs with large data sets, particularly when using the PQL method, which may be a quite good approximation in such cases. For large pedigrees it may be useful to compute the inverse of the relationship matrix using some efficient ad hoc algorithm, then to provide it as argument of the fit using the covStruct(list(precision=...))
syntax. If the precision matrix is not specified, spaMM will generally evaluate it to assess whether it should use sparse-precision methods. see sparse_precision
for further control of this computation, on another example from quantitative genetics.
## Not run: # if(requireNamespace("pedigreemm", quietly=TRUE)) { ## derived from help("pedigreemm") # p1 <- new("pedigree", sire = as.integer(c(NA,NA,1, 1,4,5)), dam = as.integer(c(NA,NA,2,NA,3,2)), label = as.character(1:6)) # A <- pedigreemm::getA(p1) ## relationship matrix # } ## => Manually-built matrix: A <- matrix(NA, ncol=6,nrow=6) A[lower.tri(A,diag=TRUE)] <- c(8,0,4,4,4,2, 8,4,0,2,5, 8,2,5,4.5, 8,5,2.5, 9,5.5, 9)/8 A <- Matrix::forceSymmetric(A,uplo = "L") colnames(A) <- rownames(A) <- 1:6 ## data simulation cholA <- chol(A) varU <- 0.4; varE <- 0.6; rep <- 20 n <- rep*6 set.seed(108) bStar <- rnorm(6, sd=sqrt(varU)) b <- crossprod(as.matrix(cholA),bStar) ID <- rep(1:6, each=rep) e0 <- rnorm(n, sd=sqrt(varE)) y <- b[ID]+e0 obs <- data.frame(y=y,IDgen=ID,IDenv=ID) ## two copies of ID for readability of GLMM results ## fits fitme(y ~ 1+ corrMatrix(1|IDgen) , corrMatrix=A,data=obs,method="REML") obs$y01 <- ifelse(y<1.3,0,1) fitme(y01 ~ 1+ corrMatrix(1|IDgen)+(1|IDenv), corrMatrix=A,data=obs, family=binomial(), method="REML") prec_mat <- solve(A) colnames(prec_mat) <- rownames(prec_mat) <- rownames(A) # important fitme(y01 ~ 1+ corrMatrix(1|IDgen)+(1|IDenv) , covStruct=list(precision=prec_mat), data=obs, family=binomial(), method="REML") ## End(Not run)
## Not run: # if(requireNamespace("pedigreemm", quietly=TRUE)) { ## derived from help("pedigreemm") # p1 <- new("pedigree", sire = as.integer(c(NA,NA,1, 1,4,5)), dam = as.integer(c(NA,NA,2,NA,3,2)), label = as.character(1:6)) # A <- pedigreemm::getA(p1) ## relationship matrix # } ## => Manually-built matrix: A <- matrix(NA, ncol=6,nrow=6) A[lower.tri(A,diag=TRUE)] <- c(8,0,4,4,4,2, 8,4,0,2,5, 8,2,5,4.5, 8,5,2.5, 9,5.5, 9)/8 A <- Matrix::forceSymmetric(A,uplo = "L") colnames(A) <- rownames(A) <- 1:6 ## data simulation cholA <- chol(A) varU <- 0.4; varE <- 0.6; rep <- 20 n <- rep*6 set.seed(108) bStar <- rnorm(6, sd=sqrt(varU)) b <- crossprod(as.matrix(cholA),bStar) ID <- rep(1:6, each=rep) e0 <- rnorm(n, sd=sqrt(varE)) y <- b[ID]+e0 obs <- data.frame(y=y,IDgen=ID,IDenv=ID) ## two copies of ID for readability of GLMM results ## fits fitme(y ~ 1+ corrMatrix(1|IDgen) , corrMatrix=A,data=obs,method="REML") obs$y01 <- ifelse(y<1.3,0,1) fitme(y01 ~ 1+ corrMatrix(1|IDgen)+(1|IDenv), corrMatrix=A,data=obs, family=binomial(), method="REML") prec_mat <- solve(A) colnames(prec_mat) <- rownames(prec_mat) <- rownames(A) # important fitme(y01 ~ 1+ corrMatrix(1|IDgen)+(1|IDenv) , covStruct=list(precision=prec_mat), data=obs, family=binomial(), method="REML") ## End(Not run)
A model can be specified for the residual-dispersion parameter of gaussian and Gamma response families. This model may or may not include random effects. The
resid.model
argument of all fitting functions is used to specify this model and to control its fit. resid.model
is either a formula (without left-hand side) for the dispersion parameter phi
of the residual error (a log link is assumed); or a list of arguments similar to those of a standard fitme
call.
The residual-dispersion model is fitted by a specific method (see Details) involving estimation of its parameters by the fit of a Gamma-response model to response values computed by the parent fitting function (i.e., the fitting function called to fit the joint models for main response and for residual dispersion). For mixed-effect residual-dispersion models, the fitme
function is used internally to fit this Gamma-response model (irrespective of the parent fitting function used, which may not be fitme
).
# 'resid.model' argument of fitting functions (fitme(), HLfit(), etc)
# 'resid.model' argument of fitting functions (fitme(), HLfit(), etc)
If resid.model
is a list, it must include a formula
element (model formula without left-hand side, as when resid.model
is only a formula
). The following additional arguments may be useful:
family |
The family is always Gamma. The default link is log. The identity link can be tried but may fail because only the log link ensures that the fitted |
fixed |
fixed values of parameters of the residual dispersion model itself. Same usage as documented in |
etaFix |
To fix some of the fixed-effect coefficients, as in the mean response, and with the same format. Note that the same effect can usually be acheived by an offset in the |
control.dist |
A list of arguments that control the computation of the distance argument of the correlation functions. Same usage as documented in |
rand.family |
A |
init , lower , upper , control
|
with same usage as documented in |
Other arguments should be ignored (see Details).
The following elements should not be specified in resid.model
, for the specified reasons:
which is constrained to be identical to the method from the parent call;
constrained to be identical to the same-named controls from the parent call;
constrained: no resid.model
for a resid.model
;
phi
of the Gamma family of the residual dispersion modelThis parameter is by default set to 1, in agreement with the theory underlying the estimation procedure for the residual model; it can be set to another value, and a resid.model
's fixed=list(phi=NA)
will even force its estimation, but this is not warranted;
constrained to NULL;
The data of the parent call are used, so they must include all the variables required for the resid.model
;
constrained: no prior weights;
constrained: use the verbose
argument of the fitting function instead to control a progress line summarizing the results of the resid.model
fit at each iteration of main loop of the parent call (see the next section of the Details).
if used, this argument may affect the fits. However, it is best ignored in practice: users would have hard time guessing good initial values, and bad ones might have unwarranted effects.
Progress reports of the fitting procedure:
Fits with a mixed-effect residual-dispersion model involve repeated “nested” fits of the latter model (each of them itself tipycally involving “double-nested” fits of a mixed-effect model with fixed random-effect parameters). This can be slow particularly when the residual-dispersion model involve spatial effects. A specific element phifit
of the verbose
vector controls screen information about progress of such fits during the full-model fit: when set to 0 (or FALSE) there is no report. For higher values a one-line message is output at the end of each “nested” call, but it may be overwritten by the next one-line message. So the ultimately visible output depends on control of overwriting. When verbose["phifit"]
is set to 1 (or TRUE) each output overwrites the previous one so the ultimately visible output is from the last “nested” call; when it is set to 2
, the final line of output of each “nested” call remains visible; when set to 3, a line of output remains visible from each “double-nested” call.
Methods: The present implementation of the Gamma-response procedure used to fit (fixed-effect) residual dispersion models is based on its exact components as detailed by Smyth et al. 2001. spaMM also implements an ML version of this REML procedure (where the leverage corrections used in the REML procedure are set to zero). Smyth et al. discuss more approximate versions of the components, considered in the early h-likelihood literature and elsewhere.
Lee and Nelder (2006) extend the REML procedure to mixed-effects residual dispersion models.
Again, spaMM also implements ML and REML versions of these procedures (using distinct “standardizing” leverages as detailed in hatvalues.HLfit
), and additionally use by default the full Laplace approximation (with observed Hessian) for fitting the Gamma-response mixed-effect model, instead of the approximation using expected Hessian considered in the h-likelihood literature.
When the residual-dispersion model includes random effects, no single likelihood objective function appears to be maximized by the joint fit of mean-response and residual dispersion models. A procedure such as numInfo
may then detect that the likelihood gradient does not vanish for all parameters. Indeed, this limitation is “relatively obvious” in original formulations relying on both marginal likelihood and restricted likelihood concepts to fit different parameters of the joint model. But this limitation is also true in the case where marginal likelihood (actually, its Laplace approximation, although the issue could persist even if exact Gamma-GLMM likelihood were used) is used in the residual-dispersion fit.
When such dispersion models are fitted, the resulting fits are embedded in the main fit object. The get_fittedPars
extractor will by default )as controlled by its argument phiPars
) include in its return value the rdisPars
element, which is the list of parameters of the residual-dispersion fit, in the same format as a get_fittedPars
value for the mean-response model (rdisPars
may also include fits of other residual-dispersion models described in resid.model
). The phi
element of the get_fittedPars
value will further contain the residual-dispersion fit itself, as a "glm"
or, when it includes random effects, as a "HLfit"
object.
Lee, Y. and Nelder, J.A. (2006), Double hierarchical generalized linear models (with discussion). Journal of the Royal Statistical Society: Series C (Applied Statistics), 55: 139-185. doi:10.1111/j.1467-9876.2006.00538.x
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006) Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
Smyth, G. K., Huele, F., and Verbyla, A. P. 2001. Exact and approximate REML for heteroscedastic regression, Stat. Modelling 1, 161–175.
data("crack") # crack data, Lee et al. 2006 chapter 11 etc hlfit <- HLfit(y~crack0+(1|specimen), family=Gamma(log), data=crack, rand.family=inverse.Gamma(log), resid.model=list(formula=~cycle+(1|specimen)) )
data("crack") # crack data, Lee et al. 2006 chapter 11 etc hlfit <- HLfit(y~crack0+(1|specimen), family=Gamma(log), data=crack, rand.family=inverse.Gamma(log), resid.model=list(formula=~cycle+(1|specimen)) )
The function pdep_effects
evaluates, and the function plot_effects
plots, partial-dependence effects.
pdep_effects
evaluates the effect of a given fixed-effect variable, as (by default, the average of) predicted values on the response scale, over the empirical distribution of all other fixed-effect variables in the data, and of inferred random effects. This can be seen as the result of an experiment where specific treatments (given values of the focal variable) are applied over all conditions defined by the other fixed effects and by the inferred random effects. Thus, apparent dependencies induced by associations between predictor variables are avoided (see Friedman, 2001, from which the name “partial dependence plot” is taken; or Hastie et al., 2009, Section 10.13.2). This also avoids biases of possible alternative ways of plotting effects. In particular, such biases occur if the response link is not identity, and if averaging is performed on the linear-predictor scale or when other variables are set to some conventional value other than its average.
pdep_effects
also compute intervals of the type defined by its intervals
argument (by default, prediction intervals) and of nominal coverage defined by the level
argument (it may make particular sense to choose a level
<0.95 to better visualize effects). By default, it returns a data frame of average values of point predictions and interval bounds for each value of the focal variable (so the intervals may briefly be described as mean prediction intervals, for want of better), but it can also return lists of all predictions.
A plot function is available for numeric or factor predictors: plot_effects
calls pdep_effects
and produces a simple plot (using only base graphic functions) of its results, including prediction bands representing the two average one-sided widths of intervals. The last section of the Examples shows how to obtain more elaborate plots including the same information using ggplot2.
If added to the plot, the raw data may appear to depart from the partial-dependence predictions, since the data are a priori affected by the associations between variables which the predictions free themselves from. An adapted plot of fit residuals may be then be more useful, and the Examples also show how it can be performed.
pdep_effects(object, focal_var, newdata = object$data, length.out = 20, focal_values=NULL, level=0.95, levels = NULL, intervals = "predVar", indiv = FALSE, ...) plot_effects(object, focal_var, newdata = object$data, focal_values=NULL, effects = NULL, xlab = focal_var, ylab = NULL, rgb.args = col2rgb("blue"), add = FALSE, ylim=NULL, ...)
pdep_effects(object, focal_var, newdata = object$data, length.out = 20, focal_values=NULL, level=0.95, levels = NULL, intervals = "predVar", indiv = FALSE, ...) plot_effects(object, focal_var, newdata = object$data, focal_values=NULL, effects = NULL, xlab = focal_var, ylab = NULL, rgb.args = col2rgb("blue"), add = FALSE, ylim=NULL, ...)
object |
An object of class |
focal_var |
Character string: the name of the predictor variable whose effect is to be represented. The variable must be numeric for |
newdata |
If non-NULL, a data frame passed to |
effects |
If non-NULL, a data frame to substitute to the one produced by default by |
xlab |
If non-NULL, a character string: X-axis label for the plot. |
ylab |
If non-NULL, a character string: Y-axis label for the plot. |
ylim |
The |
rgb.args |
Color control arguments, in the format produced by |
add |
Boolean: whether to add graphic elements of a previous plot produced by |
length.out |
Integer: for a numeric predictor variable, this controls the number of values at which predictions are evaluated. By default, predictions are made at regular intervals over the range of the predictor variable. If |
intervals , level
|
Passed to |
focal_values , levels
|
|
indiv |
Boolean: whether to return all predictions given the values of other predictors in the |
... |
Further arguments passed by |
For pdep_effects
, a nested list, or a data frame storing values of the focal_var
, average point predictions pointp
and bounds low
and up
of intervals, depending on the indiv
argument. When indiv
is TRUE
, each sublist contains vectors for pointp
, low
and up
.
For plot_effects
, the same value, returned invisibly.
J.H. Friedman (2001). Greedy Function Approximation: A Gradient Boosting Machine. Annals of Statistics 29(5):1189-1232.
J. Friedman, T. Hastie and R. Tibshirani (2009) The Elements of Statistical Learning, 2nd ed. Springer.
data("scotlip") hlcor <- HLCor(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix,family=poisson(),data=scotlip) plot_effects(hlcor,focal_var="prop.ag",ylim=c(0,max(scotlip$cases))) points(cases~prop.ag, data=scotlip, col="blue",pch=20) # Impose specific values of a numeric predictor using 'focal_values': plot_effects(hlcor, focal_var="prop.ag", focal_values=1:5) ### Adding 'partial residuals' [residuals relative to predict(<fit object>), ### but plotted relative to pdep_effects() predictions]: # One first needs predictions for actual values of the predictor variable, # provided by pdep_effects(.,length.out=0L): # pdep_points <- pdep_effects(hlcor,focal_var="prop.ag",length.out=0L) # Rename for easy prediction for each observation, and add the residuals # of the actual fit, using the default residuals() i.e. deviance ones: # rownames(pdep_points) <- pdep_points$focal_var pdep_res <- pdep_points[paste(hlcor$data$prop.ag),"pointp"] + residuals(hlcor) points(x = hlcor$data$prop.ag, y = pdep_res, col = "red", pch = 20) ## Not run: ## Plotting pdep-effects for different categories, using ggplot. library(ggplot2) data("Gryphon") tmp <- na.omit(Gryphon_df) spfit <- spaMM::fitme(TARSUS ~ BWT*sex, data = tmp) tmp$sex <- "1" pdep_1 <- pdep_effects(spfit,"BWT", newdata=tmp, level=qnorm(0.75)) # qnorm(0.75) to get the so-called 'probable error'. tmp$sex <- "2" pdep_2 <- pdep_effects(spfit,"BWT", newdata=tmp, level=qnorm(0.75)) pdep_1$sex <- "1" ; pdep_2$sex <- "2" pdep <- rbind(pdep_1,pdep_2) ggplot(pdep,aes(y = pointp , x = focal_var ,col = sex, fill=sex)) + geom_point() + geom_ribbon(aes(ymin = low, ymax = up), alpha = 0.3) + xlab("BWT") + ylab("TARSUS") ## End(Not run)
data("scotlip") hlcor <- HLCor(cases~I(prop.ag/10) +adjacency(1|gridcode)+offset(log(expec)), adjMatrix=Nmatrix,family=poisson(),data=scotlip) plot_effects(hlcor,focal_var="prop.ag",ylim=c(0,max(scotlip$cases))) points(cases~prop.ag, data=scotlip, col="blue",pch=20) # Impose specific values of a numeric predictor using 'focal_values': plot_effects(hlcor, focal_var="prop.ag", focal_values=1:5) ### Adding 'partial residuals' [residuals relative to predict(<fit object>), ### but plotted relative to pdep_effects() predictions]: # One first needs predictions for actual values of the predictor variable, # provided by pdep_effects(.,length.out=0L): # pdep_points <- pdep_effects(hlcor,focal_var="prop.ag",length.out=0L) # Rename for easy prediction for each observation, and add the residuals # of the actual fit, using the default residuals() i.e. deviance ones: # rownames(pdep_points) <- pdep_points$focal_var pdep_res <- pdep_points[paste(hlcor$data$prop.ag),"pointp"] + residuals(hlcor) points(x = hlcor$data$prop.ag, y = pdep_res, col = "red", pch = 20) ## Not run: ## Plotting pdep-effects for different categories, using ggplot. library(ggplot2) data("Gryphon") tmp <- na.omit(Gryphon_df) spfit <- spaMM::fitme(TARSUS ~ BWT*sex, data = tmp) tmp$sex <- "1" pdep_1 <- pdep_effects(spfit,"BWT", newdata=tmp, level=qnorm(0.75)) # qnorm(0.75) to get the so-called 'probable error'. tmp$sex <- "2" pdep_2 <- pdep_effects(spfit,"BWT", newdata=tmp, level=qnorm(0.75)) pdep_1$sex <- "1" ; pdep_2$sex <- "2" pdep <- rbind(pdep_1,pdep_2) ggplot(pdep,aes(y = pointp , x = focal_var ,col = sex, fill=sex)) + geom_point() + geom_ribbon(aes(ymin = low, ymax = up), alpha = 0.3) + xlab("BWT") + ylab("TARSUS") ## End(Not run)
This function provides diagnostic plots for residual errors from the mean model and for random effects. Plots for the mean models are similar
to those for GLMs. They use standardized deviance residuals as described by Lee et al. (2006, p.52). This means that plots for residual errors use the residuals provided by residuals(<fit object>, type="std_dev_res")
; and that plots for random effects likewise consider standardized values.
## S3 method for class 'HLfit' plot(x, which = c("mean", "ranef"), titles = list( meanmodel=list(outer="Mean model",devres="Deviance residuals", absdevres="|Deviance residuals|", resq="Residual quantiles", devreshist="Deviance residuals"), ranef=list(outer="Random effects and leverages",qq="Random effects Q-Q plot", levphi=expression(paste("Leverages for ",phi)), levlambda=expression(paste("Leverages for ",lambda))) ), control= list(), ask=TRUE, ...)
## S3 method for class 'HLfit' plot(x, which = c("mean", "ranef"), titles = list( meanmodel=list(outer="Mean model",devres="Deviance residuals", absdevres="|Deviance residuals|", resq="Residual quantiles", devreshist="Deviance residuals"), ranef=list(outer="Random effects and leverages",qq="Random effects Q-Q plot", levphi=expression(paste("Leverages for ",phi)), levlambda=expression(paste("Leverages for ",lambda))) ), control= list(), ask=TRUE, ...)
x |
An object of class |
which |
A vector of keywords for different types of plots.
By default, two types of plots are presented on different devices: diagnostic plots for mean values, and diagnostic plots for random effects.
Either one can be selected using this argument. Use keyword |
titles |
A list of the |
control |
A list of default options for the plots. Defaults are |
ask |
Logical; passed to |
... |
Options passed from |
In principle the standardized deviance residuals for the mean model should have a nearly Gaussian distribution hence form a nearly straight line on a Q-Q plot. However this is (trivially) not so for well-specified (nearly-)binary response data nor even for well-specified Poisson response data with moderate expectations. Hence this plot is not so useful. The DHARMa
package proposes better-behaved diagnostic plots (but the p-value that appears on one of these plots may not stand for a valid goodness-of-fit test). The current version of DHARMa
should handle spaMM
fit objects; otherwise, see https://github.com/florianhartig/DHARMa/issues/95 for how to run DHARMa
procedures on spaMM
output.
Returns the input object invisibly.
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006). Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
data("blackcap") fit <- fitme(migStatus ~ 1+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=1,nu=1,rho=1)) plot(fit)
data("blackcap") fit <- fitme(migStatus ~ 1+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=1,nu=1,rho=1)) plot(fit)
Poisson
(with a capital P) is a family
that specifies the information required to fit a Poisson generalized linear model. Differs from the base version stats::poisson
only in that it handles the zero-truncated variant, which can be specified either as Tpoisson(<link>)
or as Poisson(<link>, trunc = 0L)
. The truncated poisson with mean is defined from the un-truncated poisson with mean
, by restricting its response strictly positive value.
, where
is the probability that the response is 0.
Poisson(link = "log", trunc = -1L, LLgeneric=TRUE) Tpoisson(link="log") # <Poisson object>$linkfun(mu, mu_truncated = FALSE) # <Poisson object>$linkinv(eta, mu_truncated = FALSE)
Poisson(link = "log", trunc = -1L, LLgeneric=TRUE) Tpoisson(link="log") # <Poisson object>$linkfun(mu, mu_truncated = FALSE) # <Poisson object>$linkinv(eta, mu_truncated = FALSE)
link |
log, sqrt or identity link, specified by any of the available ways for GLM links (name, character string, one-element character vector, or object of class |
trunc |
Either |
eta , mu
|
Numeric (scalar or array). The linear predictor; and the expectation of response, truncated or not depending on |
mu_truncated |
Boolean. For |
LLgeneric |
For development purposes, not documented. |
Molas & Lesaffre (2010) developed expressions for deviance residuals for the truncated Poisson distribution, which were the ones implemented in spaMM until version 3.12.0. Later versions implement the (non-equivalent) definition as “2*(saturated_logLik - logLik)”.
predict
, when applied on an object with a truncated-response family, by default returns . The simplest way to predict
is to get the linear predictor value by
predict(.,type="link")
, and deduce using
linkinv(.)
(with default argument mu_truncated=FALSE
), since getting from
is comparatively less straightforward. The
mu.eta
member function is that of the base poisson
family, hence its mu
argument represents .
simulate
, when applied on an object with a truncated-response family, simulates the truncated family. There is currently no clean way to override this (trying to passtype="link"
to predict
will not have the intended effect).
A family
object suitable for use with glm
, as stats::
family objects.
McCullagh, P. and Nelder, J.A. (1989) Generalized Linear Models, 2nd edition. London: Chapman & Hall.
Molas M. and Lesaffre E. (2010). Hurdle models for multilevel zero-inflated data via h-likelihood. Statistics in Medicine 29: 3294-3310.
data("scotlip") logLik(glm(I(1+cases)~1,family=Tpoisson(),data=scotlip)) logLik(fitme(I(1+cases)~1+(1|id),family=Tpoisson(),fixed=list(lambda=1e-8),data=scotlip))
data("scotlip") logLik(glm(I(1+cases)~1,family=Tpoisson(),data=scotlip)) logLik(fitme(I(1+cases)~1+(1|id),family=Tpoisson(),fixed=list(lambda=1e-8),data=scotlip))
Packages implementing post-fit procedures define helper functions which may not handle spaMM's fit objects, or which have not always handled them, or which can handle them correctly only with some non-default arguments. This documentation topic gives further directions to apply some such post-fit procedures (from packages DHARMa, RLRsim, multcomp and lmerTest) to these fit objects.
For other procedures not considered here, diagnosing a failure in a debugging session may suggest a simple solution (as it did for multcomp::glht
).
For multiple comparison procedures by multcomp::glht
, one has to explicitly give the argument coef.=fixef.HLfit
(see Examples; fixef.HLfit
is the spaMM method for the generic function fixef
);
For DHARMa plots, see Details of plot.HLfit
;
For using RLRsim::RLRTSim(), see get_RLRTSim_args
.
For using lmerTest::contest() or lmerTest::anova(), see as_LMLT
.
if (requireNamespace("multcomp", quietly = TRUE)) { library(multcomp) set.seed(123) irisr <- cbind(iris,id=sample(4,replace=TRUE,size=nrow(iris))) irisfit <- fitme(Petal.Length~ Species +(1|id), data=irisr, family=Gamma(log)) summary(glht(irisfit,mcp("Species" = "Tukey"), coef.=fixef.HLfit)) }
if (requireNamespace("multcomp", quietly = TRUE)) { library(multcomp) set.seed(123) irisr <- cbind(iris,id=sample(4,replace=TRUE,size=nrow(iris))) irisfit <- fitme(Petal.Length~ Species +(1|id), data=irisr, family=Gamma(log)) summary(glht(irisfit,mcp("Species" = "Tukey"), coef.=fixef.HLfit)) }
The following functions can be used to compute point predictions and/or various measures of uncertainty associated to such predictions: *
predict
can be used for prediction of the response variable by its expected value obtained as (the inverse link transformation of) the linear predictor () and more generally for terms of the form X_n
+Z_nLv, for new design matrices X_n and Z_n.
*
Various components of prediction variances and predictions intervals can also be computed using predict
.
The get_
... functions are convenient extractors for such components; *
get_predCov_var_fix
extracts a block of a prediction covariance matrix. It was conceived for the specific purpose of computing the spatial prediction covariances between two “new” sets of geographic locations, without computing the full covariance matrix for both the new locations and the original (fitted) locations. When one of the two sets of new locations is fixed while the other varies, some expensive computations can be performed once for all sets of new locations, and be provided as the fix_X_ZAC.object
argument. The preprocess_fix_corr
extractor is designed to compute this argument.
## S3 method for class 'HLfit' predict(object, newdata = newX, newX = NULL, re.form = NULL, variances=list(), binding = FALSE, intervals = NULL, level = 0.95, blockSize = 2000L, type = "response", verbose=c(showpbar=eval(spaMM.getOption("barstyle"))), control=list(), na.action=na.omit, cluster_args=list(), ...) get_predCov_var_fix(object, newdata = NULL, fix_X_ZAC.object, fixdata, re.form = NULL, variances=list(disp=TRUE,residVar=FALSE,cov=FALSE), control=list(), ...) preprocess_fix_corr(object, fixdata, re.form = NULL, variances=list(residVar=FALSE, cov=FALSE), control=list()) get_fixefVar(...) get_predVar(..., variances=list(), which="predVar") get_residVar(...) get_respVar(...) get_intervals(..., intervals="predVar")
## S3 method for class 'HLfit' predict(object, newdata = newX, newX = NULL, re.form = NULL, variances=list(), binding = FALSE, intervals = NULL, level = 0.95, blockSize = 2000L, type = "response", verbose=c(showpbar=eval(spaMM.getOption("barstyle"))), control=list(), na.action=na.omit, cluster_args=list(), ...) get_predCov_var_fix(object, newdata = NULL, fix_X_ZAC.object, fixdata, re.form = NULL, variances=list(disp=TRUE,residVar=FALSE,cov=FALSE), control=list(), ...) preprocess_fix_corr(object, fixdata, re.form = NULL, variances=list(residVar=FALSE, cov=FALSE), control=list()) get_fixefVar(...) get_predVar(..., variances=list(), which="predVar") get_residVar(...) get_respVar(...) get_intervals(..., intervals="predVar")
object |
The return object of fitting functions |
newdata |
Either NULL, a matrix or data frame, or a numeric vector. If If |
newX |
equivalent to newdata, available for back-compatibility |
re.form |
formula for random effects to include. By default, it is NULL, in which case all random effects are included. If it is NA, no random effect is included. If it is a formula, only the random effects it contains are retained. The other variance components are removed from both point prediction and |
variances |
A list whose elements control the computation of different estimated variances.
In particular, |
intervals |
NULL or character string or vector of strings. Provides prediction intervals with nominal level |
which |
any of |
level |
Coverage of the intervals. |
binding |
If |
fixdata |
A data frame describing reference data whose covariances with variable |
fix_X_ZAC.object |
The return value of calling |
blockSize |
For data with many rows, it may be more efficient to perform some operations on slices of the data, and this gives the maximum number or rows of each slice. Further, parallelisation of computations over the slices is possible, as controlled by the |
type |
character string; The returned point predictions are on the response scale if |
control |
A list; a warning will direct you to relevant usage when needed. |
cluster_args |
Passed to |
verbose |
A vector of booleans; it single currently used element is |
na.action |
One of the functions dealing with |
... |
further arguments passed to or from other methods. For the |
See the predVar
help page for information about the different concepts of prediction variances handled by spaMM (uncertainty of point prediction vs. of response) and about options controlling their computation.
If newdata
is NULL, predict
returns the fitted responses, including random effects, from the object.
Otherwise it computes new predictions including random effects as far as possible.
For spatial random effects it constructs a correlation matrix C between new locations and locations in the original fit. Then it infers the random effects in the new locations as C (L' v (see
spaMM
for notation). For non-spatial random effects, it checks whether any group (i.e., level of a random effect) in the new data was represented in the original data, and it adds the inferred random effect for this group to the prediction for individuals in this group.
In the point prediction of the linear predictor, the unconditional expected value of is assigned to the realizations of
for unobserved levels of non-spatial random effects (it is zero in GLMMs but not for non-gaussian random effects), and the inferred value of
is assigned in all other cases. Corresponding values of
are then deduced. This computation yields the classical “BLUP” or empirical Bayes predictor in LMMs, but otherwise it may yield less well characterized predictors, where “unconditional”
may not be its expected value when the
rand.family
link is not identity.
There are cases where prediction without a newdata
argument may give results of different length than prediction with newdata=
<original data>, as for predict
. Notably, for multivariate-response fits, different subsets of lines of the data may be used for each submodel depending on the availability of all variables (including the response variable) for each submodel, and the resulting fitted values from each submodel will be used from prediction; while prediction with newdata
does not check the availability of a response variable.
Intervals computations use the relevant variance estimates plugged in a Gaussian approximation, except for the simple linear model where it uses Student's t distribution.
See Details in Tpoisson
for questions specific to truncated distributions.
For predict
, a matrix or data frame (according to the binding
argument), with optional attributes frame
, intervals
, predVar
, fixefVar
, residVar
, and/or respVar
, the last four holding one or more variance vector or covariance matrices. The further attribute fittedName
contains the binding name, if any. The frame
attribute includes information about any na.action
effect on the new data.
The get_
... extractor functions call predict
and extract from its result the attribute implied by the name of the extractor. By default, get_intervals
will return prediction intervals using predVar
.
get_predVar
with non-default which
argument has the same effect as the get_
... function whose name is implied by which
.
predVar
for information specific to prediction variances sensu lato, including the definitions of the four components of prediction variance, fixefVar
, predVar
, residVar
and respVar
, that can be requested through the variances
argument;
get_cPredVar
for a bootstrap-corrected version of get_predVar
;
residVar
for an alternative extractor for residual variances, more general than get_residVar
.
data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) predict(fitobject) #### multiple controls of prediction variances ## (1) fit with an additional random effect grouped <- cbind(blackcap,grp=c(rep(1,7),rep(2,7))) fitobject2 <- fitme(migStatus ~ 1 + (1|grp) +Matern(1|longitude+latitude), data=grouped, fixed=list(nu=4,rho=0.4,phi=0.05)) ## (2) re.form usage to remove a random effect from point prediction and variances: predict(fitobject2,re.form= ~ 1 + Matern(1|longitude+latitude)) ## (3) comparison of covariance matrices for two types of new data moregroups <- grouped[1:5,] rownames(moregroups) <- paste0("newloc",1:5) moregroups$grp <- rep(3,5) ## all new data belong to an unobserved third group cov1 <- get_predVar(fitobject2,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) moregroups$grp <- 3:7 ## all new data belong to distinct unobserved groups cov2 <- get_predVar(fitobject2,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) cov1-cov2 ## the expected off-diagonal covariance due to the common group in the first fit. ## Not run: #### Other extractors: # fix_X_ZAC.object <- preprocess_fix_corr(fitobject,fixdata=blackcap) # # ... for use in multiple calls to get_predCov_var_fix(): # get_predCov_var_fix(fitobject,newdata=blackcap[14,],fix_X_ZAC.object=fix_X_ZAC.object) #### Prediction with distinct given phi's in different locations, # as specified by a resid.model: # varphi <- cbind(blackcap,logphi=runif(14)) vphifit <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude), resid.model = list(formula=~0+offset(logphi)), data=varphi, fixed=list(nu=4,rho=0.4)) # # For respVar computation (i.e., response variance, often called prediction variance), # one then also needs to provide the variables used in 'resid.model', here 'logphi': # get_respVar(vphifit,newdata=data.frame(latitude=1,longitude=1,logphi=1)) # # For default 'predVar' computation (i.e., uncertainty in point prediction), # this is not needed: # get_predVar(vphifit,newdata=data.frame(latitude=1,longitude=1)) #### point predictions and variances with new X and Z # if(requireNamespace("rsae", quietly = TRUE)) { data("landsat", package = "rsae") fitobject <- fitme(HACorn ~ PixelsCorn + PixelsSoybeans + (1|CountyName), data=landsat[-33,]) newXandZ <- unique(data.frame(PixelsCorn=landsat$MeanPixelsCorn, PixelsSoybeans=landsat$MeanPixelsSoybeans, CountyName=landsat$CountyName)) predict(fitobject,newdata=newXandZ,variances = list(predVar=TRUE)) get_predVar(fitobject,newdata=newXandZ,variances = list(predVar=TRUE)) } ## End(Not run)
data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) predict(fitobject) #### multiple controls of prediction variances ## (1) fit with an additional random effect grouped <- cbind(blackcap,grp=c(rep(1,7),rep(2,7))) fitobject2 <- fitme(migStatus ~ 1 + (1|grp) +Matern(1|longitude+latitude), data=grouped, fixed=list(nu=4,rho=0.4,phi=0.05)) ## (2) re.form usage to remove a random effect from point prediction and variances: predict(fitobject2,re.form= ~ 1 + Matern(1|longitude+latitude)) ## (3) comparison of covariance matrices for two types of new data moregroups <- grouped[1:5,] rownames(moregroups) <- paste0("newloc",1:5) moregroups$grp <- rep(3,5) ## all new data belong to an unobserved third group cov1 <- get_predVar(fitobject2,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) moregroups$grp <- 3:7 ## all new data belong to distinct unobserved groups cov2 <- get_predVar(fitobject2,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) cov1-cov2 ## the expected off-diagonal covariance due to the common group in the first fit. ## Not run: #### Other extractors: # fix_X_ZAC.object <- preprocess_fix_corr(fitobject,fixdata=blackcap) # # ... for use in multiple calls to get_predCov_var_fix(): # get_predCov_var_fix(fitobject,newdata=blackcap[14,],fix_X_ZAC.object=fix_X_ZAC.object) #### Prediction with distinct given phi's in different locations, # as specified by a resid.model: # varphi <- cbind(blackcap,logphi=runif(14)) vphifit <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude), resid.model = list(formula=~0+offset(logphi)), data=varphi, fixed=list(nu=4,rho=0.4)) # # For respVar computation (i.e., response variance, often called prediction variance), # one then also needs to provide the variables used in 'resid.model', here 'logphi': # get_respVar(vphifit,newdata=data.frame(latitude=1,longitude=1,logphi=1)) # # For default 'predVar' computation (i.e., uncertainty in point prediction), # this is not needed: # get_predVar(vphifit,newdata=data.frame(latitude=1,longitude=1)) #### point predictions and variances with new X and Z # if(requireNamespace("rsae", quietly = TRUE)) { data("landsat", package = "rsae") fitobject <- fitme(HACorn ~ PixelsCorn + PixelsSoybeans + (1|CountyName), data=landsat[-33,]) newXandZ <- unique(data.frame(PixelsCorn=landsat$MeanPixelsCorn, PixelsSoybeans=landsat$MeanPixelsSoybeans, CountyName=landsat$CountyName)) predict(fitobject,newdata=newXandZ,variances = list(predVar=TRUE)) get_predVar(fitobject,newdata=newXandZ,variances = list(predVar=TRUE)) } ## End(Not run)
spaMM allows computation of four variance components of prediction, returned by predict
as “...Var
” attributes: predVar
, fixefVar
, residVar
, or respVar
. The phrase “prediction variance” is used inconsistently in the literature. Often it is used to denote the uncertainty in the response (therefore, including the residual variance), but spaMM follows some literature for mixed models in departing from this usage. Here, this uncertainly is called the response variance (respVar
), while prediction variance (predVar
) is used to denote the uncertainty in the linear predictor (as in Booth & Hobert, 1998; see also Jeske & Harville, 1988). The respVar
is the predVar
plus the residual variance residVar
.
Which components are returned is controlled in particular by the type
and variances
arguments of the relevant functions. variances
is a list of booleans whose possible elements either match the possible returned components: predVar
, fixefVar
, residVar
, or respVar
; or may additionally include linPred
, disp
, cov
, as_tcrossfac_list
and possibly other cryptic ones.
The predict
default value for all elements is NULL
, which jointly translate to no component being computed, equivalently to setting all elements to FALSE
. However, setting one component to TRUE
may reverse the default effect for other components. In particular, by default, component predVar
implies linPred=TRUE, disp=TRUE
and component respVar
additionally implies residVar=TRUE
; in both cases, the linPred=TRUE
default by default implies fixefVar=TRUE
. Calling for one variance may imply that some of its components are not only computed but also returned as a distinct attribute.
By default the returned components are vectors of variances (with exceptions for some type
value). To obtain covariance matrices (when applicable), set cov=TRUE
. as_tcrossfac_list=TRUE
can be used to return a list of matrices such that the
predVar
covariance matrix equals . It thus provides a representation of the
predVar
that may be useful in particular when the predVar
has large dimension, as the component s may require less memory (being possibly non-square or sparse).
residVar=TRUE
evaluates residVar
the residual variance. For families without a dispersion parameter (e.g., binomial
or poisson
), this is as given by the variance
function of the family
object (in the binomial case, it is thus the variance of a single binary draw). For families with a dispersion parameter (such as for gaussian or Gamma families, negative-binomial, beta), it is the residual variance as function of the dispersion parameter, whether this parameter is a single scalar or follows a more complex residual-dispersion model. Prior weights are however ignored (see the
residVar
etractor for the opposite feature). For the beta-binomial family, it is also the variance of a single binary draw; although this family has a residual-dispersion parameter the latter variance is not affected by it.
fixefVar=TRUE
evaluates fixefVar
, the variance due to uncertainty in fixed effects (X).
Computations implying linPred=TRUE
will take into account the variances of the linear predictor , i.e. the uncertainty in fixed effects (X
) and random effects (ZLv), for given dispersion parameters (see Details).
For fixed-effect models, the
fixefVar
calculations reduces to the linPred
one.
Computations implying disp=TRUE
additionally include the effect of uncertainty in estimates of dispersion parameters ( and
), with some limitations (see Details).
variances=list(predVar=TRUE)
, which evaluates the uncertainty of linear predictor, implies disp=TRUE
by default, meaning that it includes such effects of uncertainty in dispersion parameters on the linear predictor. variances=list(respVar=TRUE)
performs similarly but additionally includes the residual variance in the returned variance.
fixefVar
is the (co)variance of X, deduced from the asymptotic covariance matrix of
estimates.
linPred
is the prediction (co)variance of =X
+Zv (see
HLfit
Details for notation, and keep in mind that new matrices may replace the ones from the fit object when newdata
are used), by default computed for given dispersion parameters. It takes into account the joint uncertainty in estimation of and prediction of v.
In particular, for new levels of the random effects,
predVar
computation takes into account uncertainty in prediction of v for these new levels. For prediction covariance with a new Z, it matters whether a single or multiple new levels are used: see Examples.
For computations implying disp=TRUE
, prediction variance may also include a term accounting for uncertainty in and
, computed following Booth and Hobert (1998, eq. 19). This computation acheives its originally described purpose for a scalar residual variance (
) and for several random effects with scalar variances (
). This computation ignores uncertainties in spatial correlation parameters.
The (1998) formulas for the effect of uncertainty in dispersion parameters are here also applied to the variance parameters in random-coefficient terms, but with a one-time warning. Not only this is not expected to account for the uncertainty of the correlation parameter(s) of such terms, but the result will depend on the internal representation (the “square root”) of the covariance matrix, which may differ among the different fitting algorithms that may be used by spaMM.
respVar
is the sum of predVar
(pre- and post-multiplied by for models with non-identity link) and of
residVar
.
These variance calculations are approximate except for LMMs, and cannot be guaranteed to give accurate results.
Booth, J.G., Hobert, J.P. (1998) Standard errors of prediction in generalized linear mixed models. J. Am. Stat. Assoc. 93: 262-272.
Jeske, Daniel R. & Harville, David A. (1988) Prediction-interval procedures and (fixed-effects) confidence-interval procedures for mixed linear models. Communications in Statistics - Theory and Methods, 17: 1053-1087. doi:10.1080/03610928808829672
## Not run: # (but run in help("get_predVar")) data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) #### multiple controls of prediction variances # (1) fit with an additional random effect grouped <- cbind(blackcap,grp=c(rep(1,7),rep(2,7))) fitobject <- fitme(migStatus ~ 1 + (1|grp) +Matern(1|longitude+latitude), data=grouped, fixed=list(nu=4,rho=0.4,phi=0.05)) # (2) re.form usage to remove a random effect from point prediction and variances: predict(fitobject,re.form= ~ 1 + Matern(1|longitude+latitude)) # (3) comparison of covariance matrices for two types of new data moregroups <- grouped[1:5,] rownames(moregroups) <- paste0("newloc",1:5) moregroups$grp <- rep(3,5) ## all new data belong to an unobserved third group cov1 <- get_predVar(fitobject,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) moregroups$grp <- 3:7 ## all new data belong to distinct unobserved groups cov2 <- get_predVar(fitobject,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) cov1-cov2 ## the expected off-diagonal covariance due to the common group in the first fit. ## End(Not run) ## see help("get_predVar") for further examples
## Not run: # (but run in help("get_predVar")) data("blackcap") fitobject <- fitme(migStatus ~ 1 + Matern(1|longitude+latitude),data=blackcap, fixed=list(nu=4,rho=0.4,phi=0.05)) #### multiple controls of prediction variances # (1) fit with an additional random effect grouped <- cbind(blackcap,grp=c(rep(1,7),rep(2,7))) fitobject <- fitme(migStatus ~ 1 + (1|grp) +Matern(1|longitude+latitude), data=grouped, fixed=list(nu=4,rho=0.4,phi=0.05)) # (2) re.form usage to remove a random effect from point prediction and variances: predict(fitobject,re.form= ~ 1 + Matern(1|longitude+latitude)) # (3) comparison of covariance matrices for two types of new data moregroups <- grouped[1:5,] rownames(moregroups) <- paste0("newloc",1:5) moregroups$grp <- rep(3,5) ## all new data belong to an unobserved third group cov1 <- get_predVar(fitobject,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) moregroups$grp <- 3:7 ## all new data belong to distinct unobserved groups cov2 <- get_predVar(fitobject,newdata=moregroups, variances=list(linPred=TRUE,cov=TRUE)) cov1-cov2 ## the expected off-diagonal covariance due to the common group in the first fit. ## End(Not run) ## see help("get_predVar") for further examples
Generalization of R-squared based on likelihood ratios, called pseudo-R2 below, and variously attributed to Cragg & Uhler (1970), Cox & Snell (1989), Magee (1990) and some other authors (see comments in the References section). The null model used in the definition of R2 can be modified by the user.
pseudoR2(fitobject, nullform = . ~ 1, R2fun = LR2R2, rescale=FALSE, verbose=TRUE)
pseudoR2(fitobject, nullform = . ~ 1, R2fun = LR2R2, rescale=FALSE, verbose=TRUE)
fitobject |
The fitted model object, obtained as the return value of a spaMM fitting function. |
nullform |
Mean-response formula for the null model. The default value (including only an intercept) represents the traditional choice in R2 computation for linear models. Alternative formulas (including, e.g., random effects) can be specified using either the |
R2fun |
The backend function computing R2 given the fitted and null model. The default implements the pseudo-R2. For linear models, it reduces to the canonical R2 and the value adjusted as in |
rescale |
Boolean or formula, controlling whether and how to rescale R2 so that its maximum possible value is 1 (often considered for discrete-response models). If a formula, it should specify the model with maximal R2. If TRUE, rescaling is performed in a way meaningful only for binary logistic regression (see Examples for how this is implemented). |
verbose |
Boolean; whether to display various informations about the procedure (most notably, to warn about some potential problem in applying the default procedure to |
None of the R2-like computations I am aware of helps in addressing, for the general class of models handled by spaMM, a well-defined inference task (comparable to, say, formally testing goodness of fit, or measuring accuracy of prediction of new data as considered for AIC). This problem has been well-known (e.g., Magee, 1990), and the canonical R2 itself for linear models is not devoid of weaknesses from this perspective (e.g., https://stats.stackexchange.com/questions/13314/is-r2-useful-or-dangerous). As a consequence, strong statements about the properties that R2 should have are difficult to follow (and this includes the claim that it should always have maximum value 1).
Given the above problems, (1) ultimately the main reason for computing R2 may be to deal with requests by misguided reviewers; (2) no attempt has been made here to implement the wide diversity of R2-like descriptors discussed in the literature. The LR2R2
backend function implements the pseudo-R2, chosen on the basis that this is the simplest general method that makes at least as much sense as any other computation I have seen; and implementation of rescaling by maximal R2 is minimal (the examples explain some of its details). LR2R2
allows adaptation of the R2 definition for mixed-effect models, by including some random effect(s) in the null model, using the nullform
argument.
As returned by the function specified by argument R2fun
. The default function returns a numeric vector of length 2 for linear models and a single value otherwise.
Cox, D.R., Snell, E.J. (1989). The analysis of binary data (2nd ed.). Chapman and Hall.
Often cited in this context, but they barely mention the topic, in an exercise p. 208-209.
Pseudo-R2 is known to go back at least to
Cragg, J. G., & Uhler, R. S. (1970). The demand for automobiles. The Canadian Journal of Economics, 3(3), 386. doi:10.2307/133656
where they already discussed its rescaling by a maximum value, in the context of binary regression.
Magee, L. (1990) R2 Measures based on Wald and likelihood ratio joint significance tests. The American Statistician, 44, 250-253. doi:10.1080/00031305.1990.10475731
also often cited for the pseudo-R2, this paper reformulates some related descriptors and concisely reviews earlier literature.
Nagelkerke, N.J.D. (1991) A note on a general definition of the coefficient of determination. Biometrika, Vol. 78, No. 3. (Sep., 1991), pp. 691-692. doi:10.1093/biomet/78.3.691
details the properties of pseudo-R2 (including the way it “partitions” variation). Argues emphatically for its rescaling, for which it is often cited.
#### Pseudo-R2 *is* R2 for linear models: # # lmfit <- lm(sr ~ pop15+pop75+dpi+ddpi , data = LifeCycleSavings) # summary(lmfit) # Multiple R-squared = 0.3385, adjusted = 0.2797 # spfit <- fitme(sr ~ pop15+pop75+dpi+ddpi , data = LifeCycleSavings) pseudoR2(spfit) # consistent with summary(lmfit) #### Toy example of pseudo-R2 for binary data # set.seed(123) toydf <- data.frame(x=seq(50), y=sample(0:1,50,TRUE)) # ## Binary logistic regression: # binlog <- fitme(y~x, data=toydf, family=binomial()) (blR2 <- pseudoR2(binlog)) # quite low, despite the model being correct # ## Rescaling by 'maximum possible' R2 for binary logistic regression: # pseudoR2(binlog, rescale=TRUE) # # which is acheived by silently computing the maximum possible R2 value # by the following brutal but effective way: # perfbinlog <- fitme(y~I(y), data=toydf, family=binomial()) (maxblR2 <- pseudoR2(perfbinlog)) # = 0.7397... # # (this 'maximum possible' value would be modified if the null model were modified). # blR2/maxblR2 # again, rescaled value # ## Same by more general syntax: # pseudoR2(binlog, rescale=y~I(y))
#### Pseudo-R2 *is* R2 for linear models: # # lmfit <- lm(sr ~ pop15+pop75+dpi+ddpi , data = LifeCycleSavings) # summary(lmfit) # Multiple R-squared = 0.3385, adjusted = 0.2797 # spfit <- fitme(sr ~ pop15+pop75+dpi+ddpi , data = LifeCycleSavings) pseudoR2(spfit) # consistent with summary(lmfit) #### Toy example of pseudo-R2 for binary data # set.seed(123) toydf <- data.frame(x=seq(50), y=sample(0:1,50,TRUE)) # ## Binary logistic regression: # binlog <- fitme(y~x, data=toydf, family=binomial()) (blR2 <- pseudoR2(binlog)) # quite low, despite the model being correct # ## Rescaling by 'maximum possible' R2 for binary logistic regression: # pseudoR2(binlog, rescale=TRUE) # # which is acheived by silently computing the maximum possible R2 value # by the following brutal but effective way: # perfbinlog <- fitme(y~I(y), data=toydf, family=binomial()) (maxblR2 <- pseudoR2(perfbinlog)) # = 0.7397... # # (this 'maximum possible' value would be modified if the null model were modified). # blR2/maxblR2 # again, rescaled value # ## Same by more general syntax: # pseudoR2(binlog, rescale=y~I(y))
The structure of random-effect models adjustable by spaMM can generally be described by the following steps.
First, independent and identically distributed (iid) random effects u are drawn from one of the following distributions: Gaussian with zero mean, unit variance, and identity link; Beta-distributed, where
~
with mean=1/2, and var
; and with logit link
v=logit(u)
;
Gamma-distributed random effects, where ~ Gamma(shape=
1+1/,scale=1/
): see
Gamma
for allowed links and further details; and Inverse-Gamma-distributed random effects, where
~
inverse-Gamma(shape=
1+1/,rate=1/
): see
inverse.Gamma
for allowed links and further details.
Second, a transformation v(u) is applied (this defines v whose elements are still iid). By default, v=u for gaussian random effects but not necessarily for other distributions of random effects (see
Gamma
).
Third, correlated random effects are obtained as Mv, where the matrix M can describe spatial correlation between observed locations, block effects (or repeated observations in given locations), and possibly also correlations involving unobserved locations (as is often the case for autoregressive models). In most cases M is determined from the model formula, but it can also be controlled by covStruct argument. M takes the form ZL or ZAL, where Z is determined from the model formula, the optional A factor is given by the optional "AMatrices"
attribute of argument covStruct
of HLCor
(also handled by fitme
and corrHLfit
), and L can be determined from the model formula or from covStruct
. In particular:
*
Z is typically an incidence matrix: its elements are 1 if the
th observation is affected by the
th element of
ALb
, and zero otherwise.
*
For spatial random effects, L is typically the Cholesky “square root” of a correlation matrix determined by the random effect specification (e.g., Matern(...)
), or given by the covStruct
argument. This may be meaningful only for Gaussian random effects. Coefficients for each level of a random-coefficient model can also be represented as Lv where L is the “square root” of a correlation matrix.
*
If there is one response value par location, L for a spatial random effect is thus a square matrix whose dimension is the number of observations. Alternatively, several observations may be taken in the same location, and a matrix Z (automatically constructed) tells which element of Lv affects each observation. The linear predictor then contains a term of the form ZLv, where dim(Z)
is (number of observations,number of locations).
*
in IMRF
random effects (IMRF for Interpolated Markov Random Fields), the realized random effects in response locations are defined as linear combinations ALv of random effects Lv in distinct locations. In that case the dimension of L is the number of such distinct locations, an automatically constructed A matrix maps them to the observed locations, and Z again maps them to possibly repeated observations in observed locations.
By default, fitting functions in spaMM
check the rank of the design matrix for fixed effects, as stats::lm
or stats::glm
do (but not, say, nlme::lme
). This computation can be quite long.
To save time when fitting different models with the same fixed-effect terms to the same data,
the result of the check can be extracted from a return object by get_rankinfo()
,
and can be provided as argument control.HLfit$rankinfo
to another fit. Alternatively, the check will not be performed if
control.HLfit$rankinfo
is set to NA
.
get_rankinfo(object)
get_rankinfo(object)
object |
An object of class |
The check is performed by a call to qr()
methods for either dense or sparse matrices.
If the design matrix is singular, a set of columns from the design matrix that define a non-singular matrix is identified. Note that different sets may be identified by sparse- and dense-matrix qr
methods.
A list with elements rank
, whichcols
(a set of columns that define a non-singular matrix), and method
(identifying the algorithm used).
## Data preparation # Singular matrix from ?Matrix::qr : singX <- cbind(int = 1, b1=rep(1:0, each=3), b2=rep(0:1, each=3), c1=rep(c(1,0,0), 2), c2=rep(c(0,1,0), 2), c3=rep(c(0,0,1),2)) rownames(singX) <- paste0("r", seq_len(nrow(singX))) donn <- as.data.frame(singX) set.seed(123) donn$y <- runif(6) fitlm <- fitme(y~int+ b1+b2+c1+c2+c3,data=donn) get_rankinfo(fitlm)
## Data preparation # Singular matrix from ?Matrix::qr : singX <- cbind(int = 1, b1=rep(1:0, each=3), b2=rep(0:1, each=3), c1=rep(c(1,0,0), 2), c2=rep(c(0,1,0), 2), c3=rep(c(0,0,1),2)) rownames(singX) <- paste0("r", seq_len(nrow(singX))) donn <- as.data.frame(singX) set.seed(123) donn$y <- runif(6) fitlm <- fitme(y~int+ b1+b2+c1+c2+c3,data=donn) get_rankinfo(fitlm)
register_cF
registers the name of a new corrFamily constructor so that it can be used as the keyword of a random effect in a formula (as in y ~ 1 + ARp()
). unregister_cF
cancels this.
register_cF(corrFamilies = NULL, reset = FALSE) unregister_cF(corrFamilies)
register_cF(corrFamilies = NULL, reset = FALSE) unregister_cF(corrFamilies)
corrFamilies |
NULL, or character vector of names of corrFamily constructors. |
reset |
Boolean. Set it to |
No value; operates through side-effects on internal variables.
ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from 'stats' package myARp <- ARp # defines 'new' corrFamily from built-in one # Now, this would not yet work: # fitme(lh ~ 1 + myARp(1|time), data=ts, method="REML") # but this works if we first register "myARp" register_cF("myARp") # registers it fitme(lh ~ 1 + myARp(1|time), data=ts, method="REML") # # same as # fitme(lh ~ 1 + corrFamily(1|time), data=ts, method="REML", covStruct=list(corrFamily=myARp())) # # showing it's possible not to register myARp, # although this has limitations (see Details in help("corrFamily")). ## Specifying arguments of the corrFamily constructor: fitme(lh ~ 1 + myARp(1|time, p=3), data=ts, method="REML") # # same as # fitme(lh ~ 1 + corrFamily(1|time), data=ts, method="REML", covStruct=list(corrFamily=ARp(p=3))) unregister_cF("myARp") # Tidy things before leaving.
ts <- data.frame(lh=lh,time=seq(48)) ## using 'lh' data from 'stats' package myARp <- ARp # defines 'new' corrFamily from built-in one # Now, this would not yet work: # fitme(lh ~ 1 + myARp(1|time), data=ts, method="REML") # but this works if we first register "myARp" register_cF("myARp") # registers it fitme(lh ~ 1 + myARp(1|time), data=ts, method="REML") # # same as # fitme(lh ~ 1 + corrFamily(1|time), data=ts, method="REML", covStruct=list(corrFamily=myARp())) # # showing it's possible not to register myARp, # although this has limitations (see Details in help("corrFamily")). ## Specifying arguments of the corrFamily constructor: fitme(lh ~ 1 + myARp(1|time, p=3), data=ts, method="REML") # # same as # fitme(lh ~ 1 + corrFamily(1|time), data=ts, method="REML", covStruct=list(corrFamily=ARp(p=3))) unregister_cF("myARp") # Tidy things before leaving.
The resid.model
argument of fitting functions can be used to specify a model for a residual-dispersion parameter of various response families, that is, either
(1) the parameter of the gaussian and Gamma GLM families;
(2a) the dispersion parameter of some other GLM families, such as the shape parameter of the negbin1 and negbin2 families; or
(2b) the dispersion parameter of some other (non-GLM) response families, such as the precision parameter of the beta response family.
This documentation is more specifically for case (2). Case (1) is more specifically documented as phi-resid.model
.
In case (2) the model for the dispersion parameter is constrained as a fixed-effect model, of the form
dispersion parameter = exp(
X+offset
),
and specified using the standard formula
syntax. Random effects cannot be included, in contrast to dispersion models for case (1).
# 'resid.model' argument of fitme() and fitmv()
# 'resid.model' argument of fitme() and fitmv()
The resid.model
for case (2) is simply a formula (without left-hand side) for the logarithm of the dispersion parameter. Fixed values can be specified through the
rdisPars
element of the fixed
argument in the fitme
call (or through the fixed
argument of each submodel of a fitmv
call). Likewise, initial values can be specified through the init
argument.
In case (2) a fixed “heteroscedastic” model can also be specified directly through the family specification, e.g., family=negbin1(shape=<vector>) where the vector has the length of the response vector, but this may not be suitable if the model is to be used for prediction purposes (where the residual-dispersion model should be specified in such a way that one can “predict” new dispersion values from it).
The design matrix for the specified model is internally rescaled to avoid numerical problems. That means that there is no need to rescale the predictor variable, even if it tends to take large (cf ‘population’ variable in the Examples) of small values (this is also true for fixed-effect predictors of the mean-response model).
The fit results for the residual model are accessible through the summary and various extractors. In particular, the get_fittedPars
extractor will by default include in its return value the rdisPars
element, which is here the vector of fitted coefficients.
residVar(., which="fam_parm")
will return the vector of fitted values of the dispersion parameter.
data("scotlip") if (spaMM.getOption("example_maxtime")>3) { (toyfit <- fitme(cases~1+(1|id),family=negbin1(), data=scotlip, resid.model = ~ population)) # => This toy example is a bit challenging to fit because the data set is small and # individual-level variation is here described both by a random effect # and by a two-parameter negbin1 residual variation. The fit might often stop # at a local maximum of the logLik in such cases (although there is no evidence # that this is presently the case). }
data("scotlip") if (spaMM.getOption("example_maxtime")>3) { (toyfit <- fitme(cases~1+(1|id),family=negbin1(), data=scotlip, resid.model = ~ population)) # => This toy example is a bit challenging to fit because the data set is small and # individual-level variation is here described both by a random effect # and by a two-parameter negbin1 residual variation. The fit might often stop # at a local maximum of the logLik in such cases (although there is no evidence # that this is presently the case). }
Extracts several types of residuals from an object of class HLfit
. Note that the default type ("deviance"
) of returned residuals differs from the default (response residuals) of equivalent functions in base R.
## S3 method for class 'HLfit' residuals(object, type = c("deviance", "pearson", "response", "working", "std_dev_res"), force=FALSE, ...)
## S3 method for class 'HLfit' residuals(object, type = c("deviance", "pearson", "response", "working", "std_dev_res"), force=FALSE, ...)
object |
An object of class |
type |
The type of residuals which should be returned. See Details for additional information. |
force |
Boolean: to force recomputation of the |
... |
For consistency with the generic. |
The four types "deviance"
(default), "pearson"
, "response"
are "working"
are, for GLM families, the same that are returned by residuals.glm
. "working"
residuals may be returned only for fixed-effect models. "deviance"
residuals are the signed square root of those returned by dev_resids
when there are no prior weights.
In the presence of prior weights, what the standard extractors do is often a matter of confusion and spaMM has not always been consistent with them. For a gaussian-response GLM (see Examples) stats::deviance.lm
calls weighted.residuals()
which returns unscaled deviance residuals weighted by prior weights. Unscaled deviance residuals are defined in McCullagh and Nelder 1989, p. 34 and depend on the response values and fitted values but not on the canonical parameter, and prior weights are not considered.
weighted.residuals()
still ignores but accounts for prior weights. This means that different
residuals(<glm>)
and deviance(<glm>)
will be returned for equivalent fits with different parametrizations of the residual variance (as produced by glm(., family=gaussian, weights=rep(2,nrow<data>))
versus the glm
call without weights). residuals(<HLfit object>,"deviance")
and deviance(<HLfit object>,"deviance")
are consistent with this behavior. By contrast, dev_resids(<HLfit object>)
always return the unscaled deviance residuals by default.
Following Lee et al. (2006, p.52), the standardized deviance residuals returned for type="std_dev_res"
are defined as the deviance residuals divided by , where the deviance residuals are defined as for a GLM,
is the dispersion parameter of the response family (a vector of values, for heteroscedastic cases), and
is a vector of leverages given by
hatvalues(., type="std")
(see hatvalues
for details about these specific standardizing leverages).
Some definitions must be extended for non-GLM response families. In the latter case, the deviance residuals are as defined in Details of llm.fit
(there is no concept of unscaled residuals here, nor indeed of scaled ones since the residual dispersion parameter is not generally a scale factor, but the returned deviance residuals for non-GLMs are analogous to the scaled ones for GLMs as they depend on residual dispersion). "std_dev_res"
residuals are defined from them as shown above for GLM response families, with the additional convention that (since the family's own residual dispersion parameter already enters in the definition of deviance residuals for non-GLM families). Pearson residuals and response residuals are defined as in
stats:::residuals.glm
. The "working"
residuals are defined for each response as where clik is the conditional likelihood.
A vector of residuals
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006). Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
data("wafers") fit <- fitme(y ~X1+(1|batch) ,data=wafers, init=list(phi=NaN)) # : this 'init' # implies that standardized deviance residuals are saved in the # fit result, allowing the following comparison: r1 <- residuals(fit, type="std_dev_res") # gets stored value r2 <- residuals(fit, type="std_dev_res", force=TRUE) # forced recomputation if (diff(range(r1-r2))>1e-14) stop() ##### ## Not run: glmfit <- glm(I(y/1000)~X1, family=gaussian(), data=wafers) deviance(glmfit) # 3... (a) sum(residuals(glmfit)^2) # 3... (b) # Same model, with different parametrization of residual variance glmfit2 <- glm(I(y/1000)~X1, family=gaussian(), data=wafers, weights=rep(2,198)) deviance(glmfit2) # 6... (c) sum(residuals(glmfit2)^2) # 6... (d) # Same comparison but for HLfit objects: spfit <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers) deviance(spfit) # 3... (e) sum(residuals(spfit)^2) # 3... (f) sum(dev_resids(spfit)) # 3... spfit2 <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers, prior.weights=rep(2,198)) deviance(spfit2) # 6... (g) ~ (c,d) # post v4.2.0 sum(residuals(spfit2)^2) # 6... (h) ~ (c,d) sum(dev_resids(spfit2)) # 3... # Unscaled residuals should no depend on arbitrarily fixed residual variance: spfit3 <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers, fixed=list(phi=2), prior.weights=rep(2,198)) deviance(spfit3) # 6... (i) ~ (g) sum(residuals(spfit3)^2) # 6... (k) ~ (h) sum(dev_resids(spfit3)) # 3... ## End(Not run)
data("wafers") fit <- fitme(y ~X1+(1|batch) ,data=wafers, init=list(phi=NaN)) # : this 'init' # implies that standardized deviance residuals are saved in the # fit result, allowing the following comparison: r1 <- residuals(fit, type="std_dev_res") # gets stored value r2 <- residuals(fit, type="std_dev_res", force=TRUE) # forced recomputation if (diff(range(r1-r2))>1e-14) stop() ##### ## Not run: glmfit <- glm(I(y/1000)~X1, family=gaussian(), data=wafers) deviance(glmfit) # 3... (a) sum(residuals(glmfit)^2) # 3... (b) # Same model, with different parametrization of residual variance glmfit2 <- glm(I(y/1000)~X1, family=gaussian(), data=wafers, weights=rep(2,198)) deviance(glmfit2) # 6... (c) sum(residuals(glmfit2)^2) # 6... (d) # Same comparison but for HLfit objects: spfit <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers) deviance(spfit) # 3... (e) sum(residuals(spfit)^2) # 3... (f) sum(dev_resids(spfit)) # 3... spfit2 <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers, prior.weights=rep(2,198)) deviance(spfit2) # 6... (g) ~ (c,d) # post v4.2.0 sum(residuals(spfit2)^2) # 6... (h) ~ (c,d) sum(dev_resids(spfit2)) # 3... # Unscaled residuals should no depend on arbitrarily fixed residual variance: spfit3 <- fitme(I(y/1000)~X1, family=gaussian(), data=wafers, fixed=list(phi=2), prior.weights=rep(2,198)) deviance(spfit3) # 6... (i) ~ (g) sum(residuals(spfit3)^2) # 6... (k) ~ (h) sum(dev_resids(spfit3)) # 3... ## End(Not run)
Extracts from a fit object the residual variance or, depending on the which
argument, a family dispersion parameter phi (which is generally not the residual variance itself except for gaussian-response models without prior weights), or a vector of values of the dispersion parameter, or further information about the residual variance model.
For gaussian and Gamma response families, the return values for which = "var"
and "phi"
include prior weights, if any.
residVar(object, which = "var", submodel = NULL, newdata = NULL)
residVar(object, which = "var", submodel = NULL, newdata = NULL)
object |
An object of class |
which |
Character: |
submodel |
integer: the index of a submodel, if |
newdata |
Either NULL, a matrix or data frame, or a numeric vector. See |
which="var"
(default) and "phi"
always return a vector of residual variances (or, alternatively, phi values) of length determined by the newdata
and submodel
arguments. which="fit"
returns an object of class HLfit
, glm
, or a single scalar depending on the residual dispersion model (which="fit"
is the option to be used to extract the scalar phi value).which="fam_parm"
returns either NULL (for families without such a parameter), a vector (if a resid.model
was specified for relevant families), a single scalar (relevant families, without resid.model
), or a list of such objects (for multivariate-response models).
Other which
values return an object of class family
or formula
as expected.
get_residVar
is a alternative extractor of residual variances with different features inherited from get_predVar
. In particular, it is more suited for computing the residual variances of new realizations of a fitted model, not accounting for prior weights used in fitting the model (basic examples of using the IsoriX package provide a context where this is the appropriate design decision). By contrast, residVar
aims to account for prior weights.
# data preparation: simulated trivial life-history data set.seed(123) nind <- 20L u <- rnorm(nind) lfh <- data.frame( id=seq_len(nind), id2=seq_len(nind), feco= rpois(nind, lambda = exp(1+u)), growth=rgamma(nind,shape=1/0.2, scale=0.2*exp(1+u)) # mean=exp(1+u), var= 0.2*mean^2 ) # multivariate-response fit fitlfh <- fitmv(submodels=list(list(feco ~ 1+(1|id), family=poisson()), list(growth ~ 1+(1|id), family=Gamma(log))), data=lfh) # residVar(fitlfh) residVar(fitlfh, which="phi") # shows fixed phi=1 for Poisson responses residVar(fitlfh, submodel=2) residVar(fitlfh, which="family", submodel=2) residVar(fitlfh, which="formula", submodel=2) residVar(fitlfh, which="fit", submodel=2) # Fit here characterized by a single scalar
# data preparation: simulated trivial life-history data set.seed(123) nind <- 20L u <- rnorm(nind) lfh <- data.frame( id=seq_len(nind), id2=seq_len(nind), feco= rpois(nind, lambda = exp(1+u)), growth=rgamma(nind,shape=1/0.2, scale=0.2*exp(1+u)) # mean=exp(1+u), var= 0.2*mean^2 ) # multivariate-response fit fitlfh <- fitmv(submodels=list(list(feco ~ 1+(1|id), family=poisson()), list(growth ~ 1+(1|id), family=Gamma(log))), data=lfh) # residVar(fitlfh) residVar(fitlfh, which="phi") # shows fixed phi=1 for Poisson responses residVar(fitlfh, submodel=2) residVar(fitlfh, which="family", submodel=2) residVar(fitlfh, which="formula", submodel=2) residVar(fitlfh, which="fit", submodel=2) # Fit here characterized by a single scalar
Data from a salamander mating experiment discussed by McCullagh and Nelder (1989, Ch. 14). Twenty males and twenty females from two populations (Rough Butt and Whiteside) were each paired with 6 individuals from their own or from the other population. The experiments were later published by Arnold et al. (1996).
data("salamander")
data("salamander")
The data frame includes 360 observations on the following variables:
Index of the female;
Index of the male;
Whether the pair successfully mated or not;
Population of origin of female;
Population of origin of male;
Interaction term between TypeF
and TypeM
;
A factor with levels Summer
and Fall
;
Index of experiment
The data frame was borrowed from the HGLMMM
package (Molas and Lesaffre, 2011), version 0.1.2.
Arnold, S.J., Verrell, P.A., and Tilley S.G. (1996) The evolution of asymmetry in sexual isolation: a model and a test case. Evolution 50, 1024-1033.
McCullagh, P. and Nelder, J.A. (1989). Generalized Linear Models, 2nd edition. London: Chapman & Hall.
Molas, M., Lesaffre, E. (2011) Hierarchical Generalized Linear Models: The R Package HGLMMM. Journal of Statistical Software 39, 1-20.
data("salamander") ## Not run: HLfit(cbind(Mate,1-Mate)~TypeF+TypeM+TypeF*TypeM+(1|Female)+(1|Male), family=binomial(),data=salamander,method="ML") # equivalent fo using fitme(), but here a bit faster ## End(Not run)
data("salamander") ## Not run: HLfit(cbind(Mate,1-Mate)~TypeF+TypeM+TypeF*TypeM+(1|Female)+(1|Male), family=binomial(),data=salamander,method="ML") # equivalent fo using fitme(), but here a bit faster ## End(Not run)
This data set provides counts of lip cancer diagnoses made in Scottish districts from 1975 to 1980, and additional information relative to these data from Clayton and Kaldor (1987) and Breslow and Clayton (1993). The data set contains (for each district) counts of disease events and estimates of the fraction of the population involved in outdoor industry (agriculture, fishing, and forestry) which exposes it to sunlight.
data("scotlip")
actually loads a data frame,
scotlip
, and an adjacency matrix, Nmatrix
, between 56 Scottish districts,
as given by Clayton and Kaldor (1987, Table 1).
data("scotlip")
data("scotlip")
The data frame includes 56 observations on the following 7 variables:
alternative district identifier.
numeric district identifier (1 to 56).
district name.
number of lip cancer cases diagnosed 1975 - 1980.
total person years at risk 1975 - 1980.
percent of the population engaged in outdoor industry.
offsets considered by Breslow and Clayton (1993, Table 6, 'Exp' variable)
The rows are ordered according to gridcode
, so that they match the rows of Nmatrix
.
Clayton D, Kaldor J (1987). Empirical Bayes estimates of age-standardized relative risks for use in disease mapping. Biometrics, 43: 671 - 681.
Breslow, NE, Clayton, DG. (1993). Approximate Inference in Generalized Linear Mixed Models. Journal of the American Statistical Association: 88 9-25.
data("scotlip") fitme(cases~I(log(expec)), data=scotlip, adjMatrix=Nmatrix, family=poisson) ## see 'help(autoregressive)' for additional examples involving 'scotlip'.
data("scotlip") fitme(cases~I(log(expec)), data=scotlip, adjMatrix=Nmatrix, family=poisson) ## see 'help(autoregressive)' for additional examples involving 'scotlip'.
These convenient masks can be added to maps of (parts of) the world to mask map information for these areas.
However, many other tools may be available since this documentation was conceived. See e.g. the rnaturalearth package,
used to provide a sea mask in an example for filled.mapMM
data("seaMask") data("landMask") # data("worldcountries") # deprecated and removed # data("oceanmask") # deprecated and removed
data("seaMask") data("landMask") # data("worldcountries") # deprecated and removed # data("oceanmask") # deprecated and removed
seaMask
and landMask
are data frames with two variables, x
and y
for longitude and latitude.
Its contents are suitable for use with polypath
: they define different polygones, each separated by a row of NA
s.
worldcountries
and oceanmask
were sp::SpatialPolygonsDataFrame
objects previously included in spaMM (see Details for replacement). Such objects were useful for creating land masks for different geographical projections.
The removed objects worldcountries
and oceanmask
were suitable for plots involving geographical projections not available through map
, and more generally for raster plots. A land mask could be produced out of worldcountries
by filling the countries, as by fill="black"
in the code for country.layer
in the Examples in https://gitlab.mbb.univ-montp2.fr/francois/spamm-ref/-/blob/master/vignettePlus/example_raster.html. These objects may now be available through the same web page, but a better place to look for the same functionality is the IsoriX
package (objects CountryBorders
and OceanMask
).
seaMask
and landMask
were created from the world map in the maps package.
polypath
requires polygons, while map(interior=FALSE,plot=FALSE)
returns small segments. landMask
is the result of reconnecting the segments into full coastlines of all land blocks.
https://gitlab.mbb.univ-montp2.fr/francois/spamm-ref/-/blob/master/vignettePlus/example_raster.html for access to, and use of worldcountries
and oceanmask
; https://cran.r-project.org/package=IsoriX for replacement CountryBorders
and OceanMask
for these objects.
## Predicting behaviour for a land bird: simplified fit for illustration data("blackcap") bfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=0.5537,phi=1.376e-05,rho=0.0544740,nu=0.6286311)) ## the plot itself, with a sea mask, ## and an ad hoc 'pointmask' to see better the predictions on small islands # def_pointmask <- function(xy,r=1,npts=12) { theta <- 2*pi/npts *seq(npts) hexas <- lapply(seq(nrow(xy)), function(li){ p <- as.numeric(xy[li,]) hexa <- cbind(x=p[1]+r*cos(theta),y=p[2]+r*sin(theta)) rbind(rep(NA,2),hexa) ## initial NA before each polygon }) do.call(rbind,hexas) } ll <- blackcap[,c("longitude","latitude")] pointmask <- def_pointmask(ll[c(2,4,5,6,7),],r=0.8) ## small islands only # if (spaMM.getOption("example_maxtime")>1) { data("seaMask") filled.mapMM(bfit,add.map=TRUE, plot.title=title(main="Inferred migration propensity of blackcaps", xlab="longitude",ylab="latitude"), decorations=quote(points(pred[,coordinates],cex=1,pch="+")), plot.axes=quote({axis(1);axis(2); polypath(rbind(seaMask,pointmask),border=FALSE, col="grey", rule="evenodd") })) }
## Predicting behaviour for a land bird: simplified fit for illustration data("blackcap") bfit <- fitme(migStatus ~ means+ Matern(1|longitude+latitude),data=blackcap, fixed=list(lambda=0.5537,phi=1.376e-05,rho=0.0544740,nu=0.6286311)) ## the plot itself, with a sea mask, ## and an ad hoc 'pointmask' to see better the predictions on small islands # def_pointmask <- function(xy,r=1,npts=12) { theta <- 2*pi/npts *seq(npts) hexas <- lapply(seq(nrow(xy)), function(li){ p <- as.numeric(xy[li,]) hexa <- cbind(x=p[1]+r*cos(theta),y=p[2]+r*sin(theta)) rbind(rep(NA,2),hexa) ## initial NA before each polygon }) do.call(rbind,hexas) } ll <- blackcap[,c("longitude","latitude")] pointmask <- def_pointmask(ll[c(2,4,5,6,7),],r=0.8) ## small islands only # if (spaMM.getOption("example_maxtime")>1) { data("seaMask") filled.mapMM(bfit,add.map=TRUE, plot.title=title(main="Inferred migration propensity of blackcaps", xlab="longitude",ylab="latitude"), decorations=quote(points(pred[,coordinates],cex=1,pch="+")), plot.axes=quote({axis(1);axis(2); polypath(rbind(seaMask,pointmask),border=FALSE, col="grey", rule="evenodd") })) }
A classic toy data set, “from research conducted by microbiologist Dr P. Whitney of Surrey University. A batch of tiny seeds is brushed onto a plate covered with a certain extract at a given dilution. The numbers of germinated and ungerminated seeds are subsequently counted” (Crowder, 1978). Two seed types and two extracts are here considered in a 2x2 factorial design.
data("seeds")
data("seeds")
The data frame includes 21 observations on the following variables:
Factor for replication;
Seed type, a factor with two levels O73 and O75;
Root extract, a factor with two levels Bean and Cucumber;
Number of seeds that germinated;
Total number of seeds tested
Crowder (1978), Table 3.
Crowder, M.J., 1978. Beta-binomial anova for proportions. Appl. Statist., 27, 34-37.
Y. Lee and J. A. Nelder. 1996. Hierarchical generalized linear models (with discussion). J. R. Statist. Soc. B, 58: 619-678.
# An extended quasi-likelihood (EQL) fit as considered by Lee & Nelder (1996): data("seeds") fitme(cbind(r,n-r)~seed*extract+(1|plate),family=binomial(), rand.family=Beta(), method="EQL-", # see help("method") for difference with "EQL+" method data=seeds)
# An extended quasi-likelihood (EQL) fit as considered by Lee & Nelder (1996): data("seeds") fitme(cbind(r,n-r)~seed*extract+(1|plate),family=binomial(), rand.family=Beta(), method="EQL-", # see help("method") for difference with "EQL+" method data=seeds)
A few steps of fitting can be parallelized. Currently it is possible to control the use of multiple threads by OpenMP by the Eigen library. By default only one thread will be used, but this may be modified by using control.HLfit$NbThreads
in a fitting function's arguments, as in
avail_thr <- parallel::detectCores(logical=FALSE) - 1L fitme(., control.HLfit=list(NbThreads=max(avail_thr, 1L)))
This control is distinct from that of post-fit steps such as bootstraps where some parallel computations are controlled e.g. the nb_cores
argument of spaMM_boot
. In cases where post-fits computation imply refits of models (as is typical of parametric bootstraps), the two parallelizations should not be combined, and the spaMM code for post-fit operations will in principle automatically take care of this.
According to https://cran.r-project.org/doc/manuals/r-devel/R-exts.html#OpenMP-support, using openMP may decrease the precision of some computations, and may be less efficient under Windows; and according to https://eigen.tuxfamily.org/dox/TopicMultiThreading.html only a few Eigen computations will benefit from such parallelisation, mainly the dense matrix products. spaMM will suggest using parallelisation when random effects have many levels and dense-correlation methods are selected (see algebra
), that is mainly for geostatiscal models with many locations. Speed gains appear moderate, as the slowest steps are not parallelized.
From an HLfit object, simulate.HLfit
function generates new samples given the estimated fixed effects
and dispersion parameters. Simulation may be unconditional (the default, useful in many applications of parametric bootstrap), or conditional on the predicted values of random effects, or may draw from the conditional distribution of random effects given the observed response. Simulations may be run for the original sampling design (i.e., original values of fixed-effect predictor variables and of random-effect levels, including spatial locations if relevant), or for a new design as specified by the newdata
argument.
simulate4boot
is a wrapper around simulate.HLfit
that can be used to precompute the bootstrap samples to be used by spaMM_boot
or spaMM2boot
through their boot_samples
argument (and is called internally by these functions when boot_samples
is NULL).
simulate_ranef
will only simulate and return a vector of random effects, more specifically some elements of the b vector appearing in the standard form offset
+ X + Z b for the linear predictor.
## S3 method for class 'HLfit' simulate(object, nsim = 1, seed = NULL, newdata = NULL, type = "marginal", re.form, conditional = NULL, verbose = c(type=TRUE, showpbar=eval(spaMM.getOption("barstyle"))), sizes = if (is.null(newdata)) object$BinomialDen, resp_testfn = NULL, phi_type = "predict", prior.weights = if (is.null(newdata)) object$prior.weights, variances=list(), ...) ## S3 method for class 'HLfitlist' simulate(object, nsim = 1, seed = NULL, newdata = object[[1]]$data, sizes = NULL, ...) simulate4boot(object, nsim, seed=NULL, resp_testfn=NULL, type="marginal", showpbar=eval(spaMM.getOption("barstyle")), ...) simulate_ranef(object, which=NULL, newdata = NULL, nsim = 1L)
## S3 method for class 'HLfit' simulate(object, nsim = 1, seed = NULL, newdata = NULL, type = "marginal", re.form, conditional = NULL, verbose = c(type=TRUE, showpbar=eval(spaMM.getOption("barstyle"))), sizes = if (is.null(newdata)) object$BinomialDen, resp_testfn = NULL, phi_type = "predict", prior.weights = if (is.null(newdata)) object$prior.weights, variances=list(), ...) ## S3 method for class 'HLfitlist' simulate(object, nsim = 1, seed = NULL, newdata = object[[1]]$data, sizes = NULL, ...) simulate4boot(object, nsim, seed=NULL, resp_testfn=NULL, type="marginal", showpbar=eval(spaMM.getOption("barstyle")), ...) simulate_ranef(object, which=NULL, newdata = NULL, nsim = 1L)
object |
The return object of HLfit or similar function. |
nsim |
number of response vectors to simulate. Defaults to '1'. |
seed |
A seed for |
newdata |
A data frame closely matching the original data, except that response values are not needed. May provide new values of fixed predictor variables, new spatial locations, new individuals within a block, or new values of the LHS in random-effect terms of the form |
prior.weights |
Prior weights that may be substituted to those of the original fit, with the same effect on the residual variance.
See Details for the definition of the default when |
sizes |
A vector of sample sizes to simulate in the case of a binomial or |
re.form |
formula for random effects to condition on. Default behaviour depends on the |
type |
character string specifying which uncertainties are taken into account in the linear predictor and notably in the random effect terms. Whatever the |
conditional |
Obsolete and will be deprecated. Boolean; TRUE and FALSE are equivalent to |
verbose |
Either a single boolean (which determines |
resp_testfn |
NULL, or a function that tests a condition which simulated samples should satisfy. This function takes a response vector as argument and return a boolean (TRUE indicating that the sample satisfies the condition). |
phi_type |
Character string, either |
variances |
Used when |
... |
For |
which |
Integer, or integer vector: the random effect(s) (indexed as ordered as in the model formula) to be simulated. If NULL, all of them are simulated. |
showpbar |
Controls display of progress bar. See |
type="predVar"
accounts for the uncertainty of the linear predictor, by drawing new values of the predictor in a multivariate gaussian distribution with mean and covariance matrix of prediction. In this case, the user has to provide a variances
argument, passed to predict
, which controls what goes into this covariance matrix. For example, with variances=list(linPred=TRUE,disp=TRUE)
), the covariance matrix takes into account the joint uncertainty in the fixed-effect coefficients and of any random effects given the response and the point estimates of dispersion and correlation parameters ("linPred"
variance component), and in addition accounts for uncertainty in the dispersion parameters (effect of "disp"
variance component as further described in predict.HLfit
). The total simulation variance is then the response variance. Uncertainty in correlation parameters (such a parameters of the Matern family) is not taken into account. The "linPred"
uncertainty is known exactly in LMMs, and otherwise approximated as a Gaussian distribution with mean vector and covariance matrix given as per the Laplace approximation.
type="(ranef|response)"
can be viewed as a special version of type="predVar"
wherevariances=list(linPred=TRUE,disp=FALSE)
) and only the uncertainty in the random effects is taken into account.
A full discussion of the merits of the different type
s is beyond the scope of this documentation, but these different types may not all be useful. type="marginal"
is typically used for computation of confidence intervals by parametric bootstrap methods. type="residual"
is used by get_cPredVar
for its evaluation of a bias term. The other type
s may be used to simulate the uncertainty in the random effects, conditionally on the data, and may therefore be more akin to the computation of prediction intervals conditionally on an (unknown but inferred) realization of the random effects. But these should presumably not be used in a bootstrap computation of such intervals, as this would represent a double accounting of the uncertainty that the bootstrap aims to quantify.
There are cases where simulation without a newdata
argument may give results of different length than simulation with newdata=
<original data>, as for predict
.
When newdata
are provided but new values of prior.weights
or sizes
are missing, new values of these missing arguments are guessed, and warnings may be issued depending on the kind of guess made for response families dependent on such arguments. The prior.weights
values used in the fit are re-used without warning when such values were identical (generally, unit) for all response values, and labelled as such in the object$prior.weights
. Unit weights will be used otherwise, with a warning. Unit binomial sizes will be used, with a warning, whenever there are newdata
.
simulate.HLfit
returns a vector (if nsim=1) or a matrix with nsim
columns, each containing simulated responses (or simulated random effects, for simulated_ranef()
). For multivariate-response simulations, an nobs
attribute gives the number of responses for each submodel if no resp_testfn
was applied.
simulate4boot
returns a list with elements
the result of simulate.HLfit
as a matrix with nsim
columns;
the state of .Random.seed
at the beginning of the sample simulation.
The simulate.HLfitlist
method returns a list of simulated responses.
data("Loaloa") HLC <- HLCor(cbind(npos,ntot-npos)~Matern(1|longitude+latitude), data=Loaloa,family=binomial(), ranPars=list(lambda=1,nu=0.5,rho=1/0.7)) simulate(HLC,nsim=2) ## Structured dispersion model data("wafers") hl <- HLfit(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) simulate(hl,type="marginal",phi_type="simulate",nsim=2) simulate_ranef(hl,nsim=2)
data("Loaloa") HLC <- HLCor(cbind(npos,ntot-npos)~Matern(1|longitude+latitude), data=Loaloa,family=binomial(), ranPars=list(lambda=1,nu=0.5,rho=1/0.7)) simulate(HLC,nsim=2) ## Structured dispersion model data("wafers") hl <- HLfit(y ~X1+X2+X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) simulate(hl,type="marginal",phi_type="simulate",nsim=2) simulate_ranef(hl,nsim=2)
Fits a range of mixed-effect models, including those with spatially correlated random effects. The random effects are either Gaussian (which defines GLMMs), or other distributions (which defines the wider class of hierarchical GLMs), or simply absent (which makes a LM or GLM). Multivariate-response models can be fitted by the fitmv
function. Other models can be fitted by fitme
. Also available are previously conceived fitting functions HLfit
(sometimes faster, for non-spatial models), HLCor
(sometimes faster, for conditional-autoregressive models and fixed-correlation models), and corrHLfit
(now of lesser interest). A variety of post-fit procedures are available for prediction, simulation and testing (see, e.g., fixedLRT
, simulate
and predict
).
A variety of special syntaxes for fixed effects, such as poly
, splines::
ns
or bs
, or lmDiallel::GCA
, may be handled natively although some might not be fully handled by post-fit procedures such as predict
. poly
is fully handled. lmDiallel::GCA
cannot be due to its inherent limitations, but see X.GCA
for a more functional alternative for diallel/multi-membership fixed-effect terms. Note that packages implementing these terms must be attached to the search list as ::
will not be properly understood in a formula
.
Both maximum likelihood (ML) and restricted likelihood (REML) can be used for linear mixed models, and extensions of these methods using Laplace approximations are used for non-Gaussian random response. Several variants of these methods discussed in the literature are included (see Details in HLfit
), the most notable of which may be “PQL/L” for binary-response GLMMs (see Example for arabidopsis
data). PQL methods implemented in spaMM are closer to (RE)ML methods than those implemented in MASS::glmmPQL
.
The standard response families gaussian
, binomial
, poisson
, and Gamma
are handled, as well as negative binomial (see negbin1
and negbin2
), beta (beta_resp
), beta-binomial (betabin
), zero-truncated poisson and negative binomial and Conway-Maxwell-Poisson response (see Tpoisson
, Tnegbin
and COMPoisson
). A multi
family look-alike is also available for multinomial
response, with some constraints.
The variance parameter of residual error is denoted (
phi
): this is the residual variance for gaussian response, but for Gamma-distributed response, the residual variance is where
is expected response. A (possibly mixed-effects) linear predictor for
, modeling heteroscedasticity, can be considered (see Examples).
The package fits models including several nested or crossed random effects, including autocorrelated ones. An interface is being developed allowing users to implement their own parametric correlation models (see corrFamily
), beyond the following ones which are built in spaMM:
* geostatistical (Matern
, Cauchy
),
* interpolated Markov Random Fields (IMRF
, MaternIMRFa
),
* autoregressive time-series (AR1
, ARp
, ARMA
),
* conditional autoregressive as specified by an adjacency
matrix,
* pairwise interactions with individual-level random effects, such as diallel experiments (diallel
),
* or any fixed correlation matrix (corrMatrix
).
GLMMs and HGLMs are fit via Laplace approximations for (1) the marginal likelihood with respect to random effects and (2) the restricted likelihood (as in REML), i.e. the likelihood of random effect parameters given the fixed effect estimates. All handled models can be formulated in terms of a linear predictor of the traditional form offset
+ X + Z b, where X is the design matrix of fixed effects,
(
beta
) is a vector of fixed-effect coefficients, Z is a “design matrix” for the random effects (which is instead denoted M=ZAL elsewhere in the package documentation), and b a vector of random effect values. The general structure of Mb is described in random-effects
.
Gaussian and non-gaussian random effects can be fitted. Different gaussian random-effect terms are handled, with the following effects:
* (1|<RHS>), for non-autocorrelated random effects as in lme4; * (<LHS>|<RHS>), for random-coefficient terms as in lme4, *and additional terms depending on the <LHS> type* (further detailed below); * (<LHS> || <RHS>) is interpreted as in lme4: any such term is immediately converted to ( (1|<RHS>) + (0+<LHS>|<RHS>) ). It should be counted as two random effects for all purposes (e.g., for fixing the variances of the random effects). However, this syntax is useless when the LHS includes a factor (see help('lme4::expandDoubleVerts')). * <prefix>(1|<RHS>), to specify autocorrelated random effects, e.g. Matern(1|long+lat). * <prefix>(<LHS>|<RHS>), where the <LHS> can be used to alter the autocorrelated random effect as detailed below.
Different LHS types of gaussian (<LHS>|<RHS>)
random-effect terms are handled, with the following effects:
* <logical> (TRUE/FALSE): affects only responses for which <LHS> is TRUE. * <factor built from a logical>: same a <logical> case; * <factor not built from a logical>: random-coefficient term as in lme4; * 0 + <factor not built from a logical>: same but contrasts are not used; * factors specified by the mv(...) expression, generate random-coefficient terms specific to multivariate-response models fitted by fitmv() (see help("mv")). 0 + mv(...) has the expected effect of not using contrasts; * <numeric> (but not '0+<numeric>'): random-coefficient term as in lme4, with 2*2 covariance matrix of effects on Intercept and slope; * 0 + <numeric>: no Intercept so no covariance matrix (random-slope-only term);
The '0 + <numeric>' effect is achieved by direct control of the elements of the incidence matrix Z through the <LHS>
term: for numeric z
, such elements are multiplied by z
values, and thus provide a variance of order O(z
squared).
If one wishes to fit uncorrelated group-specific random-effects with distinct variances for different groups or for different response variables, three syntaxes are thus possible. The most general, suitable for fitting several variances (see GxE
for an example), is to fit a (0 + <factor>| <RHS>) random-coefficient term with correlation(s) fixed to 0. Alternatively, one can define numeric (0|1) variables for each group (as as.numeric(<boolean for given group membership>)
), and use each of them in a 0 + <numeric>
LHS (so that the variance of each such random effect is zero for response not belonging to the given group). See lev2bool
for various ways of specifying such indicator variables for several levels.
Gaussian <prefix>(<LHS not 1>|<RHS>)
random-effect terms may be handled, with two main cases depending on the LHS type, motivated by the following example: independent Matérn effects can be fitted for males and females by using the syntax Matern(male|.) + Matern(female|.)
, where male
and female
are TRUE/FALSE (or a factor with TRUE/FALSE levels). In contrast to a (male|.)
term, no random-coefficient correlation matrix is fitted. However, for some other types of RHS, one can fit composite random effects combining a random-coefficient correlation matrix and the correlation model defined by the “prefix”. This combination is defined in composite-ranef
. This leads to the following distinction:
* The terms are *not* composite random effects when the non-‘1
’ LHS type is boolean or factor-from-boolean, a just illustrated, but also 0+<numeric>
: for example, Matern(0+<numeric>|.)
represents an autocorrelated random-slope (only) term or, equivalently, a direct specification of heteroscedasticity of the Matérn random effect.
* By contrast, Matern(<numeric>|.)
implies estimating a random-coefficient covariance matrix and thus defines a composite random effects, as does an LHS that is a factor constructed from numeric or character levels.
Composite random effects can be fitted in principle for all “prefixes”, including for <corrFamily>
terms. In practice, this functionality has been checked for Matern
, corrMatrix
, AR1
and the ARp
-corrFamily term. In these terms, the <.> %in% <.>
form of nested random effect is allowed.
The syntax (z-1|.)
, for numeric z
only, can also be used to fit some heteroscedastic non-Gaussian random effects. For example, a Gamma random-effect term (wei-1|block)
specifies an heteroscedastic Gamma random effect with constant mean 1 and variance
wei^2
, where
is still the estimated variance parameter. See Details of
negbin
for a possible application. Here, this effect is not implemented through direct control of Z (multiplying the elements of an incidence matrix Z by wei
), as this would have a different effect on the distribution of the random effect term. (z|.)
is not defined for non-Gaussian random effects. It could mean that a correlation structure between random intercepts and random slopes for (say) Gamma-distributed random effects is considered, but such correlation structures are not well-specified by their correlation matrix.
spaMM
was initially published by François Rousset and Jean-Baptiste Ferdy, and is continually developed by F. Rousset and tested by Alexandre Courtiol.
Lee, Y., Nelder, J. A. and Pawitan, Y. (2006). Generalized linear models with random effects: unified analysis via h-likelihood. Chapman & Hall: London.
Rousset F., Ferdy, J.-B. (2014) Testing environmental and genetic effects in the presence of spatial autocorrelation. Ecography, 37: 781-790. doi:10.1111/ecog.00566
See the test
directory of the package for many additional examples of spaMM usage beyond those from the formal documentation.
See fitme
for multivariate-response models.
data("wafers") data("scotlip") ## loads 'scotlip' data frame, but also 'Nmatrix' ## Linear model fitme(y ~ X1, data=wafers) ## GLM fitme(y ~ X1, family=Gamma(log), data=wafers) fitme(cases ~ I(log(population)), data=scotlip, family=poisson) ## Non-spatial GLMMs fitme(y ~ 1+(1|batch), family=Gamma(log), data=wafers) fitme(cases ~ 1+(1|gridcode), data=scotlip, family=poisson) # # Random-slope model (mind the output!) fitme(y~X1+(X2|batch),data=wafers, method="REML") ## Spatial, conditional-autoregressive GLMM if (spaMM.getOption("example_maxtime")>2) { fitme(cases ~ I(log(population))+adjacency(1|gridcode), data=scotlip, family=poisson, adjMatrix=Nmatrix) # with adjacency matrix provided by data("scotlip") } # see ?adjacency for more details on these models ## Spatial, geostatistical GLMM: # see e.g. examples in ?fitme, ?corrHLfit, ?Loaloa, or ?arabidopsis; # see examples in ?Matern for group-specific spatial effects. ## Hierachical GLMs with non-gaussian random effects data("salamander") if (spaMM.getOption("example_maxtime")>1) { # both gaussian and non-gaussian random effects fitme(cbind(Mate,1-Mate)~1+(1|Female)+(1|Male),family=binomial(), rand.family=list(gaussian(),Beta(logit)),data=salamander) # Random effect of Male nested in that of Female: fitme(cbind(Mate,1-Mate)~1+(1|Female/Male), family=binomial(),rand.family=Beta(logit),data=salamander) # [ also allowed is cbind(Mate,1-Mate)~1+(1|Female)+(1|Male %in% Female) ] } ## Modelling residual variance ( = structured-dispersion models) # GLM response, fixed effects for residual variance fitme( y ~ 1,family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) # # GLMM response, and mixed effects for residual variance if (spaMM.getOption("example_maxtime")>1.5) { fitme(y ~ 1+(1|batch),family=Gamma(log), resid.model = ~ 1+(1|batch) ,data=wafers) }
data("wafers") data("scotlip") ## loads 'scotlip' data frame, but also 'Nmatrix' ## Linear model fitme(y ~ X1, data=wafers) ## GLM fitme(y ~ X1, family=Gamma(log), data=wafers) fitme(cases ~ I(log(population)), data=scotlip, family=poisson) ## Non-spatial GLMMs fitme(y ~ 1+(1|batch), family=Gamma(log), data=wafers) fitme(cases ~ 1+(1|gridcode), data=scotlip, family=poisson) # # Random-slope model (mind the output!) fitme(y~X1+(X2|batch),data=wafers, method="REML") ## Spatial, conditional-autoregressive GLMM if (spaMM.getOption("example_maxtime")>2) { fitme(cases ~ I(log(population))+adjacency(1|gridcode), data=scotlip, family=poisson, adjMatrix=Nmatrix) # with adjacency matrix provided by data("scotlip") } # see ?adjacency for more details on these models ## Spatial, geostatistical GLMM: # see e.g. examples in ?fitme, ?corrHLfit, ?Loaloa, or ?arabidopsis; # see examples in ?Matern for group-specific spatial effects. ## Hierachical GLMs with non-gaussian random effects data("salamander") if (spaMM.getOption("example_maxtime")>1) { # both gaussian and non-gaussian random effects fitme(cbind(Mate,1-Mate)~1+(1|Female)+(1|Male),family=binomial(), rand.family=list(gaussian(),Beta(logit)),data=salamander) # Random effect of Male nested in that of Female: fitme(cbind(Mate,1-Mate)~1+(1|Female/Male), family=binomial(),rand.family=Beta(logit),data=salamander) # [ also allowed is cbind(Mate,1-Mate)~1+(1|Female)+(1|Male %in% Female) ] } ## Modelling residual variance ( = structured-dispersion models) # GLM response, fixed effects for residual variance fitme( y ~ 1,family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) # # GLMM response, and mixed effects for residual variance if (spaMM.getOption("example_maxtime")>1.5) { fitme(y ~ 1+(1|batch),family=Gamma(log), resid.model = ~ 1+(1|batch) ,data=wafers) }
spaMM_boot
simulates samples from a fit object inheriting from class "HLfit"
, as produced by spaMM's fitting functions, and applies a given function to each simulated sample. Parallelization is supported (see Details).
spaMM2boot
is similar except that it assumes that the original model is refitted on the simulated data, and the given function is applied to the refitted model, and the value is in a format directly usable as input for boot::boot.ci
.
Both of these functions can be used to apply standard parametric bootstrap procedures. spaMM_boot
is suitable for more diverse applications, e.g. to fit by one model some samples simulated under another model (see Example).
spaMM_boot(object, simuland, nsim, nb_cores=NULL, seed=NULL, resp_testfn=NULL, control.foreach=list(), debug. = FALSE, type, fit_env=NULL, cluster_args=NULL, showpbar= eval(spaMM.getOption("barstyle")), boot_samples=NULL, ...) spaMM2boot(object, statFUN, nsim, nb_cores=NULL, seed=NULL, resp_testfn=NULL, control.foreach=list(), debug. = FALSE, type="marginal", fit_env=NULL, cluster_args=NULL, showpbar= eval(spaMM.getOption("barstyle")), boot_samples=NULL, ...)
spaMM_boot(object, simuland, nsim, nb_cores=NULL, seed=NULL, resp_testfn=NULL, control.foreach=list(), debug. = FALSE, type, fit_env=NULL, cluster_args=NULL, showpbar= eval(spaMM.getOption("barstyle")), boot_samples=NULL, ...) spaMM2boot(object, statFUN, nsim, nb_cores=NULL, seed=NULL, resp_testfn=NULL, control.foreach=list(), debug. = FALSE, type="marginal", fit_env=NULL, cluster_args=NULL, showpbar= eval(spaMM.getOption("barstyle")), boot_samples=NULL, ...)
object |
The fit object to simulate from. |
simuland |
The function to apply to each simulated sample. See Details for requirements of this function. |
statFUN |
The function to apply to each fit object for each simulated sample. See Details for requirements of this function. |
nsim |
Number of samples to simulate and analyze. |
nb_cores |
Number of cores to use for parallel computation. The default is |
seed |
Passed to |
resp_testfn |
Passed to |
control.foreach |
list of control arguments for |
debug. |
Boolean (or integer, interpreted as boolean). For debugging purposes, given that |
type |
Character: passed to |
fit_env |
An environment or list containing variables necessary to evaluate |
cluster_args |
|
showpbar |
Controls display of progress bar. See |
boot_samples |
NULL, or precomputed bootstrap samples from the fitted model, provided as a matrix with one column per bootstrap replicate (the format of the result of |
... |
Further arguments passed to the |
The simuland
function must take as first argument a vector of response values, and may have other arguments including ‘...’. When required, these additional arguments must be passed through the ‘...’ arguments of spaMM_boot
. Variables needed to evaluate them must be available from within the simuland
function or otherwise provided as elements of fit_env
.
The statFUN
function must take as first argument (named refit
) a fit object, and may have other arguments including ‘...’ handled as for simuland
.
spaMM_boot
handles parallel backends with different features. pbapply::pbapply
has a very simple interface (essentially equivalent to apply
) and provides progress bars, but (in version 1.4.0, at least) does not have efficient load-balancing. doSNOW
also provides a progress bar and allows more efficient load-balancing, but its requires foreach
. foreach
handles errors differently from pbapply
(which will simply stop if fitting a model to a bootstrap replicate fails): see the foreach
documentation.
spaMM_boot
calls simulate.HLfit
on the fit object
and applies simuland
on each column of the matrix returned by this call.
simulate.HLfit
uses the type
argument, which must be explicitly provided.
spaMM_boot
returns a list, with the following element(s) (unless debug.
is TRUE
):
nsim
return values in the format returned either by apply
or parallel::parApply
or by foreach::`%dopar%`
as controlled by control.foreach$.combine
(which is here "rbind"
by default).
(absent in the case the boot_samples
argument was used to provide the new response values but not the RNGstate
) the state of .Random.seed
at the beginning of the sample simulation
.
spaMM2boot
returns a list suitable for use by boot.ci
, with elements:
nsim
return values of the simulated statistic (in matrix format).
nsim
return the value of statFUN
from the original fit.
The simulation type ("parametric"
).
nsim
the state of .Random.seed
at the beginning of the sample simulation
.
(other elements of an object of class boot
are currently not included.)
if (spaMM.getOption("example_maxtime")>7) { data("blackcap") # Generate fits of null and full models: lrt <- fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap) # The 'simuland' argument: myfun <- function(y, what=NULL, lrt, ...) { data <- lrt$fullfit$data data$migStatus <- y ## replaces original response (! more complicated for binomial fits) full_call <- getCall(lrt$fullfit) ## call for full fit full_call$data <- data res <- eval(full_call) ## fits the full model on the simulated response if (!is.null(what)) res <- eval(what)(res=res) ## post-process the fit return(res) ## the fit, or anything produced by evaluating 'what' } # where the 'what' argument (not required) of myfun() allows one to control # what the function returns without redefining the function. # Call myfun() with no 'what' argument: returns a list of fits spaMM_boot(lrt$nullfit, simuland = myfun, nsim=1, lrt=lrt, type ="marginal")[["bootreps"]] # Return only a model coefficient for each fit: spaMM_boot(lrt$nullfit, simuland = myfun, nsim=7, what=quote(function(res) fixef(res)[2L]), lrt=lrt, type ="marginal")[["bootreps"]] ## Not run: # Parametric bootstrap by spaMM2boot() and spaMM_boot(): boot.ci_info <- spaMM2boot(lrt$nullfit, statFUN = function(refit) fixef(refit)[1], nsim=99, type ="marginal") boot::boot.ci(boot.ci_info, , type=c("basic","perc","norm")) nullfit <- lrt$nullfit boot_t <- spaMM_boot(lrt$nullfit, simuland = function(y, nullfit) { refit <- update_resp(nullfit, y) fixef(refit)[1] }, nsim=99, type ="marginal", nullfit=nullfit)$bootreps boot::boot.ci(list(R = length(boot_t), sim="parametric"), t0=fixef(nullfit)[1], t= t(boot_t), type=c("basic","perc","norm")) ## End(Not run) }
if (spaMM.getOption("example_maxtime")>7) { data("blackcap") # Generate fits of null and full models: lrt <- fixedLRT(null.formula=migStatus ~ 1 + Matern(1|longitude+latitude), formula=migStatus ~ means + Matern(1|longitude+latitude), method='ML',data=blackcap) # The 'simuland' argument: myfun <- function(y, what=NULL, lrt, ...) { data <- lrt$fullfit$data data$migStatus <- y ## replaces original response (! more complicated for binomial fits) full_call <- getCall(lrt$fullfit) ## call for full fit full_call$data <- data res <- eval(full_call) ## fits the full model on the simulated response if (!is.null(what)) res <- eval(what)(res=res) ## post-process the fit return(res) ## the fit, or anything produced by evaluating 'what' } # where the 'what' argument (not required) of myfun() allows one to control # what the function returns without redefining the function. # Call myfun() with no 'what' argument: returns a list of fits spaMM_boot(lrt$nullfit, simuland = myfun, nsim=1, lrt=lrt, type ="marginal")[["bootreps"]] # Return only a model coefficient for each fit: spaMM_boot(lrt$nullfit, simuland = myfun, nsim=7, what=quote(function(res) fixef(res)[2L]), lrt=lrt, type ="marginal")[["bootreps"]] ## Not run: # Parametric bootstrap by spaMM2boot() and spaMM_boot(): boot.ci_info <- spaMM2boot(lrt$nullfit, statFUN = function(refit) fixef(refit)[1], nsim=99, type ="marginal") boot::boot.ci(boot.ci_info, , type=c("basic","perc","norm")) nullfit <- lrt$nullfit boot_t <- spaMM_boot(lrt$nullfit, simuland = function(y, nullfit) { refit <- update_resp(nullfit, y) fixef(refit)[1] }, nsim=99, type ="marginal", nullfit=nullfit)$bootreps boot::boot.ci(list(R = length(boot_t), sim="parametric"), t0=fixef(nullfit)[1], t= t(boot_t), type=c("basic","perc","norm")) ## End(Not run) }
spaMM_glm.fit
is a stand-in replacement for glm.fit
, which can be called through glm
by using
glm(<>, method="spaMM_glm.fit")
. Input and output structure are exactly as for glm.fit
. It uses a Levenberg-Marquardt algorithm to prevent divergence of estimates. For models families such as Gamma()
(with default inverse link) where the linear predictor is constrained to be positive, if the rcdd package is installed, the function can automatically find valid starting values or else indicate that no parameter value is feasible. It also automatically provides good starting values in some cases where the base functions request them from the user (notably, for gaussian(log)
with some negative response).
spaMM_glm
is a convenient wrapper, calling glm
with default method glm.fit
, then calling method spaMM_glm.fit
, with possibly different initial values, if glm.fit
failed.
spaMM_glm.fit(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL, mustart = NULL, offset = rep(0, nobs), family = gaussian(), control = list(maxit=200), intercept = TRUE, singular.ok = TRUE) spaMM_glm(formula, family = gaussian, data, weights, subset, na.action, start = NULL, etastart, mustart, offset, control = list(...), model = TRUE, method = c("glm.fit","spaMM_glm.fit"), x = FALSE, y = TRUE, singular.ok = TRUE, contrasts = NULL, strict=FALSE, ...)
spaMM_glm.fit(x, y, weights = rep(1, nobs), start = NULL, etastart = NULL, mustart = NULL, offset = rep(0, nobs), family = gaussian(), control = list(maxit=200), intercept = TRUE, singular.ok = TRUE) spaMM_glm(formula, family = gaussian, data, weights, subset, na.action, start = NULL, etastart, mustart, offset, control = list(...), model = TRUE, method = c("glm.fit","spaMM_glm.fit"), x = FALSE, y = TRUE, singular.ok = TRUE, contrasts = NULL, strict=FALSE, ...)
All arguments except strict
are common to these functions and their stats
package equivalents, glm
and glm.fit
. Most arguments operate as for the latter functions, whose documentation is repeated below. The control
argument may operate differently.
formula |
an object of class |
family |
a description of the error distribution and link
function to be used in the model. For |
data |
an optional data frame, list or environment (or object
coercible by |
weights |
an optional vector of ‘prior weights’ to be used
in the fitting process. Should be |
subset |
an optional vector specifying a subset of observations to be used in the fitting process. |
na.action |
a function which indicates what should happen
when the data contain |
start |
starting values for the parameters in the linear predictor. |
etastart |
starting values for the linear predictor. |
mustart |
starting values for the vector of means. |
offset |
this can be used to specify an a priori known
component to be included in the linear predictor during fitting.
This should be |
control |
a list of parameters for controlling the fitting
process. This is passed to |
model |
a logical value indicating whether model frame should be included as a component of the returned value. |
method |
A 2-elements vector specifying first the method to be used by |
x , y
|
For For |
singular.ok |
logical; if |
contrasts |
an optional list. See the |
intercept |
logical. Should an intercept be included in the null model? |
strict |
logical. Whether to perform a fit by |
... |
arguments to be used to form the default
|
An object inheriting from class glm
. See glm
for details.
The source and documentation is derived in large part from those of glm.fit
.
x <- c(8.752,20.27,24.71,32.88,27.27,19.09) y <- c(5254,35.92,84.14,641.8,1.21,47.2) # glm(.) fails: (check_error <- try(glm(y~ x,data=data.frame(x,y),family=Gamma(log)), silent=TRUE)) if ( ! inherits(check_error,"try-error")) stop("glm(.) call unexpectedly succeeded") spaMM_glm(y~ x,data=data.frame(x,y),family=Gamma(log)) ## Gamma(inverse) examples x <- c(43.6,46.5,21.7,18.6,17.3,16.7) y <- c(2420,708,39.6,16.7,46.7,10.8) # glm(.) fails (can't find starting value) (check_error <- suppressWarnings(try(glm(y~ x,data=data.frame(x,y),family=Gamma()) , silent=TRUE))) if ( ! inherits(check_error,"try-error")) stop("glm(.) call unexpectedly succeeded.") if (requireNamespace("rcdd",quietly=TRUE)) { spaMM_glm(y~ x,data=data.frame(x,y),family=Gamma()) } ## A simple exponential regression with some negative response values set.seed(123) x <- seq(50) y <- exp( -0.1 * x) + rnorm(50, sd = 0.1) glm(y~ x,data=data.frame(x,y),family=gaussian(log), method="spaMM_glm.fit") # => without the 'method' argument, stats::gaussian(log)$initialize() is called # and stops on negative response values.
x <- c(8.752,20.27,24.71,32.88,27.27,19.09) y <- c(5254,35.92,84.14,641.8,1.21,47.2) # glm(.) fails: (check_error <- try(glm(y~ x,data=data.frame(x,y),family=Gamma(log)), silent=TRUE)) if ( ! inherits(check_error,"try-error")) stop("glm(.) call unexpectedly succeeded") spaMM_glm(y~ x,data=data.frame(x,y),family=Gamma(log)) ## Gamma(inverse) examples x <- c(43.6,46.5,21.7,18.6,17.3,16.7) y <- c(2420,708,39.6,16.7,46.7,10.8) # glm(.) fails (can't find starting value) (check_error <- suppressWarnings(try(glm(y~ x,data=data.frame(x,y),family=Gamma()) , silent=TRUE))) if ( ! inherits(check_error,"try-error")) stop("glm(.) call unexpectedly succeeded.") if (requireNamespace("rcdd",quietly=TRUE)) { spaMM_glm(y~ x,data=data.frame(x,y),family=Gamma()) } ## A simple exponential regression with some negative response values set.seed(123) x <- seq(50) y <- exp( -0.1 * x) + rnorm(50, sd = 0.1) glm(y~ x,data=data.frame(x,y),family=gaussian(log), method="spaMM_glm.fit") # => without the 'method' argument, stats::gaussian(log)$initialize() is called # and stops on negative response values.
input arguments are generally similar to those of glm
and (g)lmer
, in particular for the spaMM::fitme
function, with the exception of the prior.weights
argument, which is simply weights
in the other packages.
The name prior.weights
seems more consistent, since e.g. glm
returns its input weights
as output prior.weights
, while its output weights
are instead the weights in the final iteration of an iteratively weighted least-square fit.
The default likelihood target for dispersion parameters is restricted likelihood (REML estimation) for corrHLfit
and (marginal) likelihood (ML estimation) for fitme
.
Model fits may provide restricted likelihood values(ReL
) even if restricted likelihood is is not used as an objective function at any step in the analysis.
See good-practice
for advice about the proper syntax of formula
.
Computation times depend on control parameters given by spaMM.getOption("spaMM_tol")
parameters (for iterative algorithms), and spaMM.getOption("nloptr")
parameters for the default optimizer. Do not use spaMM.options()
to control them globally, unless you know what you are doing. Rather control them locally by the control.HLfit
argument to control spaMM_tol
, and by the control arguments of corrHLfit
and fitme
to control nloptr
. If nloptr$Xtol_rel
is set above 5e-06, fitme
will by default refit the fixed effects and dispersion parameters (but not other correlation parameters estimated by nloptr
) by the iterative algorithm after nloptr
convergence. Increasing nloptr$Xtol_rel
value may therefore switches the bulk of computation time from the optimizer to the iterative algorithm, and may increase or decrease computation time depending on which algorithm is faster for a given input. Use control$refit
if you wish to inhibit this, but note that by default it provides a rescue to a poor nloptr
result due to a too large Xtol_rel
.
Chambers J.M. (2008) Software for data analysis: Programming with R. Springer-Verlag New York
spaMM.colors
is the default color palette for some color plots in spaMM
.
spaMM.colors(n = 64, redshift = 1, adjustcolor_args=NULL)
spaMM.colors(n = 64, redshift = 1, adjustcolor_args=NULL)
n |
Number of color levels returned by the function. A calling graphic function with argument |
redshift |
The higher it is, the more the palette blushes.... |
adjustcolor_args |
Either NULL or a list of arguments for |
If you don't like this color palette, have a look at the various ones provided by the fields
package.
A vector giving the colors in a hexadecimal format.
## see mapMM examples
## see mapMM examples
This function is derived from filled.contour
in the graphics
package, and
this documentation is likewise heavily based on that of filled.contour
.
This function likewise produces a contour plot with the areas between the
contours filled in solid color, and a
key showing how the colors map to z values is likewise shown to the right of
the plot.
The only difference is the way the aspect ratio is determined and can be controlled (using the map.asp
parameter instead of asp
),
They thus easily provide nice-looking maps with meaningful latitude/longitude ratio (see Examples). However, this does not work well with rstudio.
spaMM.filled.contour(x = seq(0, 1, length.out = nrow(z)), y = seq(0, 1, length.out = ncol(z)), z, xrange = range(x, finite = TRUE), yrange = range(y, finite = TRUE), zrange = range(z, finite = TRUE, na.rm=TRUE), margin=1/20, levels = pretty(zrange, nlevels), nlevels = 20, color.palette = spaMM.colors, col = color.palette(length(levels) - 1), plot.title, plot.axes, key.title=NULL, key.axes=NULL, map.asp = NULL, xaxs = "i", yaxs = "i", las = 1, axes = TRUE, frame.plot = axes, ...)
spaMM.filled.contour(x = seq(0, 1, length.out = nrow(z)), y = seq(0, 1, length.out = ncol(z)), z, xrange = range(x, finite = TRUE), yrange = range(y, finite = TRUE), zrange = range(z, finite = TRUE, na.rm=TRUE), margin=1/20, levels = pretty(zrange, nlevels), nlevels = 20, color.palette = spaMM.colors, col = color.palette(length(levels) - 1), plot.title, plot.axes, key.title=NULL, key.axes=NULL, map.asp = NULL, xaxs = "i", yaxs = "i", las = 1, axes = TRUE, frame.plot = axes, ...)
x , y
|
locations of grid lines at which the values in |
z |
a numeric matrix containing the values to be plotted.. Note that
|
xrange |
x range of the plot. |
yrange |
y range of the plot. |
zrange |
z range of the plot. |
margin |
This controls how far (in relative terms) the plot extends beyond the x and y ranges
of the analyzed points, and is overriden by explicit |
levels |
a set of levels which are used to partition the range
of |
nlevels |
if |
color.palette |
a color palette function to be used to assign colors in the plot. |
col |
an explicit set of colors to be used in the plot. This argument overrides any palette function specification. There should be one less color than levels |
plot.title |
statements which add titles to the main plot. |
plot.axes |
statements which draw axes (and a |
key.title |
statements which add titles for the plot key. |
key.axes |
statements which draw axes on the plot key. This overrides the default axis. |
map.asp |
the y/x aspect ratio of the 2D plot area (not of the full figure including the scale). Default is (plotted y range)/(plotted x range) (i.e., scales for x are identical). |
xaxs |
the x axis style. The default is to use internal labeling. |
yaxs |
the y axis style. The default is to use internal labeling. |
las |
the style of labeling to be used. The default is to use horizontal labeling. |
axes , frame.plot
|
logicals indicating if axes and a box should be
drawn, as in |
... |
additional graphical parameters, currently only passed to
|
The values to be plotted can contain NA
s. Rectangles with two
or more corner values are NA
are omitted entirely: where there
is a single NA
value the triangle opposite the NA
is
omitted.
Values to be plotted can be infinite: the effect is similar to that
described for NA
values.
This returns invisibly a list with elements of the plot, the x
, y
, z
coordinates and the contour levels
.
Builds heavily on filled.contour by Ross Ihaka and R-core.
spaMM.filled.contour
uses the layout
function and so is
restricted to a full page display.
The output produced by spaMM.filled.contour
is actually a combination
of two plots; one is the filled contour and one is the legend. Two
separate coordinate systems are set up for these two plots, but they
are only used internally – once the function has returned these
coordinate systems are lost. If you want to annotate the main contour
plot, for example to add points, you can specify graphics commands in
the plot.axes
argument. See the Examples.
Cleveland, W. S. (1993) Visualizing Data. Summit, New Jersey: Hobart.
contour
, image
,
palette
; contourplot
and levelplot
from package lattice
.
spaMM.filled.contour(volcano, color.palette = spaMM.colors) # simple ## Comparing the layout with that of filled.contour: # (except that it does not always achieve the intended effect # in RStudio Plots pane). x <- 10*1:nrow(volcano) y <- 10*1:ncol(volcano) spaMM.filled.contour(x, y, volcano, color.palette = terrain.colors, plot.title = title(main = "The Topography of Maunga Whau", xlab = "Meters North", ylab = "Meters West"), plot.axes = { axis(1, seq(100, 800, by = 100)) axis(2, seq(100, 600, by = 100)) }, key.title = title(main = "Height\n(meters)"), key.axes = axis(4, seq(90, 190, by = 10))) # maybe also asp = 1 mtext(paste("spaMM.filled.contour(.) from", R.version.string), side = 1, line = 4, adj = 1, cex = .66) ## compare with filled.contour(x, y, volcano, color.palette = terrain.colors, plot.title = title(main = "The Topography of Maunga Whau", xlab = "Meters North", ylab = "Meters West"), plot.axes = { axis(1, seq(100, 800, by = 100)) axis(2, seq(100, 600, by = 100)) }, key.title = title(main = "Height\n(meters)"), key.axes = axis(4, seq(90, 190, by = 10))) # maybe also asp = 1 mtext(paste("filled.contour(.) from", R.version.string), side = 1, line = 4, adj = 1, cex = .66)
spaMM.filled.contour(volcano, color.palette = spaMM.colors) # simple ## Comparing the layout with that of filled.contour: # (except that it does not always achieve the intended effect # in RStudio Plots pane). x <- 10*1:nrow(volcano) y <- 10*1:ncol(volcano) spaMM.filled.contour(x, y, volcano, color.palette = terrain.colors, plot.title = title(main = "The Topography of Maunga Whau", xlab = "Meters North", ylab = "Meters West"), plot.axes = { axis(1, seq(100, 800, by = 100)) axis(2, seq(100, 600, by = 100)) }, key.title = title(main = "Height\n(meters)"), key.axes = axis(4, seq(90, 190, by = 10))) # maybe also asp = 1 mtext(paste("spaMM.filled.contour(.) from", R.version.string), side = 1, line = 4, adj = 1, cex = .66) ## compare with filled.contour(x, y, volcano, color.palette = terrain.colors, plot.title = title(main = "The Topography of Maunga Whau", xlab = "Meters North", ylab = "Meters West"), plot.axes = { axis(1, seq(100, 800, by = 100)) axis(2, seq(100, 600, by = 100)) }, key.title = title(main = "Height\n(meters)"), key.axes = axis(4, seq(90, 190, by = 10))) # maybe also asp = 1 mtext(paste("filled.contour(.) from", R.version.string), side = 1, line = 4, adj = 1, cex = .66)
Large matrices and other memory-expensive objects may be stored in a fit object. This function removes them in order to reduce the size of the object, particularly when stored on disk. In principle, the removed objects can be regenerated automatically when needed (e.g., for a predict()).
stripHLfit(object, ...)
stripHLfit(object, ...)
object |
The result of a fit (an object of class |
... |
Further arguments, not currently used. |
The input fit objects with some elements removed.
The effect may change without notice between versions as the efficiency of the operation is highly sensitive to implementation details.
## Not run: ## rather unconvincing example : quantitative effect is small. # measure size of saved object: saveSize <- function (object,...) { tf <- tempfile(fileext = ".RData") on.exit(unlink(tf)) save(object, file = tf,...) file.size(tf) } data("Loaloa") lfit <- fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), method="HL(0,1)", data=Loaloa, family=binomial(), fixed=list(nu=0.5,rho=1,lambda=0.5)) saveSize(lfit) pfit <- predict(lfit,newdata=Loaloa,variances=list(cov=TRUE)) # increases size! saveSize(lfit) lfit <- stripHLfit(lfit) saveSize(lfit) ## End(Not run)
## Not run: ## rather unconvincing example : quantitative effect is small. # measure size of saved object: saveSize <- function (object,...) { tf <- tempfile(fileext = ".RData") on.exit(unlink(tf)) save(object, file = tf,...) file.size(tf) } data("Loaloa") lfit <- fitme(cbind(npos,ntot-npos)~elev1+elev2+elev3+elev4+maxNDVI1+seNDVI +Matern(1|longitude+latitude), method="HL(0,1)", data=Loaloa, family=binomial(), fixed=list(nu=0.5,rho=1,lambda=0.5)) saveSize(lfit) pfit <- predict(lfit,newdata=Loaloa,variances=list(cov=TRUE)) # increases size! saveSize(lfit) lfit <- stripHLfit(lfit) saveSize(lfit) ## End(Not run)
Summary and print methods for results from HLfit or related functions. summary
may also be used as an extractor (see e.g. beta_table
).
## S3 method for class 'HLfit' summary(object, details=FALSE, max.print=100L, verbose=TRUE, ...) ## S3 method for class 'HLfitlist' summary(object, ...) ## S3 method for class 'fixedLRT' summary(object, verbose=TRUE, ...) ## S3 method for class 'HLfit' print(x,...) ## S3 method for class 'HLfitlist' print(x,...) ## S3 method for class 'fixedLRT' print(x,...)
## S3 method for class 'HLfit' summary(object, details=FALSE, max.print=100L, verbose=TRUE, ...) ## S3 method for class 'HLfitlist' summary(object, ...) ## S3 method for class 'fixedLRT' summary(object, verbose=TRUE, ...) ## S3 method for class 'HLfit' print(x,...) ## S3 method for class 'HLfitlist' print(x,...) ## S3 method for class 'fixedLRT' print(x,...)
object |
An object of class |
x |
The return object of HLfit or related functions. |
verbose |
For |
max.print |
Controls |
details |
A vector with elements controlling whether to print some obscure details. Element |
... |
further arguments passed to or from other methods. |
The random effect terms of the linear predictor are of the form ZLv. In particular, for random-coefficients models (i.e., including random-effect terms such as (z|group)
specifying a random-slope component), correlated random effects are represented as b = Lv for some matrix L, and where the elements of v are uncorrelated. In the output of the fit, the Var.
column gives the
variances of the correlated effects, b=Lv. The Corr.
column(s) give their correlation(s). If details
is TRUE, estimates and SEs of the (log) variances of the elements of v are reported as for other random effects in the Estimate
and cond.SE.
columns of the table of lambda coefficients. However, this non-default output is potentially misleading as the elements of v cannot generally be assigned to specific terms (such as intercept and slope) of the random-effect formula, and the representation of b as Lv is not unique.
The return value is a list whose elements may be subject to changes, but two of them can be considered stable, and are thus part of the API: the beta_table
and lambda_table
which are the displayed tables for the coefficients of fixed effects and random-effect variances.
## see examples of fitme() or corrHLfit() usage
## see examples of fitme() or corrHLfit() usage
update
and update_resp
will update and (by default) re-fit a model. They do this mostly by extracting the call stored in the object, updating the call and evaluating that call. Using update(<fit>)
is a risky programming style (see Details). update_formulas(<mv fit>, ...)
can update formulas from a fitmv
fit as well as the single formula of a fit by the other fitting functions.
update_resp
handles a new response vector as produced by simulate
.
## S3 method for class 'HLfit' update(object, formula., ..., evaluate = TRUE) update_resp(object, newresp, ..., evaluate = TRUE) update_formulas(object, formula., ...) # <fit object>$respName[s] : see Details.
## S3 method for class 'HLfit' update(object, formula., ..., evaluate = TRUE) update_resp(object, newresp, ..., evaluate = TRUE) update_formulas(object, formula., ...) # <fit object>$respName[s] : see Details.
object |
A return object from an HLfit call. |
formula. |
A standard |
newresp |
New response vector. |
... |
Additional arguments to the call, or arguments with changed values. Use name |
evaluate |
If TRUE, evaluate the new call else return the call. |
Controlling response updating: Updating the data may be tricky when the response specified in a formula is not simply the name of a variable in the data. For example,
if the response was specified as I(foo^2)
the variable foo
is not what simulate.HLfit
will simulate, so foo
should not be updated with such simulation results, yet this is what should be updated in the data
. For some time spaMM has handled such cases by using an alternative way to provide updated response information, but this has some limitations. So spaMM now update the data after checking that this is correct, which the consequence that when response updating is needed (notably, for bootstrap procedures), the response should preferably be specified as the name of a variable in the data, rather than a more complicated expression.
However, in some cases, dynamic evaluation of the response variable may be helpful. For example, for bootstrapping hurdle models, the zero-truncated response may be specified as I(count[presence>0] <- NA; count)
(where both the zero-truncated count
and binary presence
variables are both updated by the bootstrap simulation). In that case the names of the two variables to be updated is provided by setting (say)<fit object>$respNames <- c("presence", "count")
for an hurdle model fit as a bivariate-response model, with first submodel for presence/absence, and second submodel for zero-truncated response. A full example is developed in the “Gentle introduction” to spaMM (
https://gitlab.mbb.univ-montp2.fr/francois/spamm-ref/-/blob/master/vignettePlus/spaMMintro.pdf).
Alternatively for univariate-response fits, use <fit object>$respName <- "count"
Controlling formula updating: Early versions of spaMM's update
method relied on stats::update.formula
whose results endorse stats
's (sometimes annoying) convention that a formula without an explicit intercept term actually includes an intercept. spaMM::update.HLfit
was then defined to avoid this problem. Formula updates should still be carefully checked, as getting them perfect has not been on the priority list.
Various post-fit functions from base R may use update.formula
directly, rather than using automatic method selection for update
. update.formula
is not itself a generic, which leads to the following problem. To make update.formula()
work on multivariate-response fits, one would like to be able to redefine it as a generic, with an HLfit
method that would perform what update_formulas
does, but such a redefinition appears to be forbidden in a package distributed on CRAN. Instead it is suggested to define a new generic spaMM::update
, which could have a spaMM::update.formula
as a method (possibly itself a generic). This would be of limited interest as the new spaMM::update.formula
would be visible to spaMM::update
but not to stats::update
, and thus the post-fit functions from base R would still not use this method.
Safe updating: update(<fit>, ...)
, as a general rule, is tricky. update
methods are easily affected in a non-transparent way by changes in variables used in the original call. For example
foo <- rep(1,10) m <- lm(rnorm(10)~1, weights=foo) rm(foo) update(m, .~.) # Error
To avoid such problems, spaMM tries to avoid references to variables in the global environment, by enforcing that the data are explicitly provided to the fitting functions by the data
argument, and that any variable used in the prior.weights
argument is in the data.
Bugs can also result when calling update
on a fit produced within some function, say function somefn
calling fitme(data=mydata,...)
, as e.g. update(<fit>)
will then seek a global variable mydata
that may differ from the fitted mydata
which was local to somefn
.
update.formula(object)
returns an object of the same nature as formula(object)
. The other functions and methods return an HLfit fit of the same type as the input object, or a call object, depending on the evaluate
value. Warning: The object returned by update_resp
cannot be used safely for further programming, for the reason explained in the Details section.
data("wafers") ## First the fit to be updated: wFit <- HLfit(y ~X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) newresp <- simulate(wFit) update_resp(wFit,newresp=newresp) # For estimates given by Lee et al., Appl. Stochastic Models Bus. Ind. (2011) 27: 315-328: # Refit with given beta or/and phi values: betavals <- c(5.55,0.08,-0.14,-0.21,-0.08,-0.09,-0.09) # reconstruct fitted phi value from predictor for log(phi) Xphi <- with(wafers,cbind(1,X3,X3^2)) ## design matrix phifit <- exp(Xphi %*% c(-2.90,0.1,0.95)) upd_wafers <- wafers designX <- get_matrix(wFit) upd_wafers$off_b <- designX %*% betavals update(wFit,formula.= . ~ offset(off_b)+(1|batch), data=upd_wafers, ranFix=list(lambda=exp(-3.67),phi=phifit)) ## There are subtlety in performing REML fits of constrained models, ## illustrated by the fact that the following fit does not recover ## the original likelihood values, because dispersion parameters are ## estimated but the REML correction changes with the formula: upd_wafers$off_f <- designX %*% fixef(wFit) ## = predict(wFit,re.form=NA,type="link") update(wFit,formula.= . ~ offset(off_f)+(1|batch), data=upd_wafers) # ## To maintain the original REML correction, Consider instead update(wFit,formula.= . ~ offset(off_f)+(1|batch), data=upd_wafers, REMLformula=formula(wFit)) ## recover original p_v and p_bv ## Alternatively, show original wFit as differences from betavals: update(wFit,formula.= . ~ . +offset(off_f), data=upd_wafers)
data("wafers") ## First the fit to be updated: wFit <- HLfit(y ~X1*X3+X2*X3+I(X2^2)+(1|batch),family=Gamma(log), resid.model = ~ X3+I(X3^2) ,data=wafers) newresp <- simulate(wFit) update_resp(wFit,newresp=newresp) # For estimates given by Lee et al., Appl. Stochastic Models Bus. Ind. (2011) 27: 315-328: # Refit with given beta or/and phi values: betavals <- c(5.55,0.08,-0.14,-0.21,-0.08,-0.09,-0.09) # reconstruct fitted phi value from predictor for log(phi) Xphi <- with(wafers,cbind(1,X3,X3^2)) ## design matrix phifit <- exp(Xphi %*% c(-2.90,0.1,0.95)) upd_wafers <- wafers designX <- get_matrix(wFit) upd_wafers$off_b <- designX %*% betavals update(wFit,formula.= . ~ offset(off_b)+(1|batch), data=upd_wafers, ranFix=list(lambda=exp(-3.67),phi=phifit)) ## There are subtlety in performing REML fits of constrained models, ## illustrated by the fact that the following fit does not recover ## the original likelihood values, because dispersion parameters are ## estimated but the REML correction changes with the formula: upd_wafers$off_f <- designX %*% fixef(wFit) ## = predict(wFit,re.form=NA,type="link") update(wFit,formula.= . ~ offset(off_f)+(1|batch), data=upd_wafers) # ## To maintain the original REML correction, Consider instead update(wFit,formula.= . ~ offset(off_f)+(1|batch), data=upd_wafers, REMLformula=formula(wFit)) ## recover original p_v and p_bv ## Alternatively, show original wFit as differences from betavals: update(wFit,formula.= . ~ . +offset(off_f), data=upd_wafers)
summary(<fit object>)$beta_table
returns the table of fixed-effect coefficients as it is printed by summary
, including standard errors and t-values.
vcov
returns the variance-covariance matrix of the fixed-effects coefficients (cf See Also for related computations involving random effects).
Corr
by default returns correlation matrices of random effects (though see Details for user-defined correlation models).
VarCorr
returns (co)variance parameters of random effects, and optionally the residual variance(s), from a fit object, in a data frame format roughly consistent with the method of objects of class "lme"
, in particular including columns with consistent names for easier extraction. One may have to consult the summary of the object to check the meaning of the contents of this data frame (e.g., of 'variance' coefficients for non-gaussian random effects). Other extractors to consider are get_ranPars
and get_inits_from_fit
, the latter providing parameters in a form suitable for initializing a fit.
The covariance matrix of residuals of a fit can be obtained as a block of the hat matrix
(get_matrix(., which="hat_matrix")
).
This is (as other covariances matrices above) a matrix of expected values, generally assuming that the fitted model is correct and that its parameters are “well” estimated, and should not to be confused with the computation of diagnostic correlations among inferred residuals of a fit.
## S3 method for class 'HLfit' vcov(object, ...) ## S3 method for class 'HLfit' VarCorr(x, sigma = 1, add_residVars=TRUE, verbose=TRUE, ...) Corr(object, A=TRUE, cov2cor.=TRUE, ...)
## S3 method for class 'HLfit' vcov(object, ...) ## S3 method for class 'HLfit' VarCorr(x, sigma = 1, add_residVars=TRUE, verbose=TRUE, ...) Corr(object, A=TRUE, cov2cor.=TRUE, ...)
object , x
|
A fitted model object, inheriting from class |
add_residVars |
Boolean; whether to include residual variance information in the returned table. |
sigma |
ignored argument, included for consistency with the generic function. |
A |
Boolean: Whether to return the correlation matrix described by the AL matrix product, when there is an A matrix (as for |
cov2cor. |
Boolean: Whether to convert covariance matrices to correlation matrices (see Details). |
verbose |
Boolean: Whether to print some notes. |
... |
Other arguments that may be needed by some method. |
Any matrix returned by the Corr
extractor is by default the unconditional correlation matrix of a vector “ALv” of random effects (as defined in random-effects
).
But it may also be an heteroscedastic matrix for some random effects if cov2cor.
is set to FALSE. In particular, the IMRF
and MaternIMRFa
models are by default defined in terms of the inverse of an heteroscedastic covariance matrix (with tcrossprod
factor L), and of a A matrix of weights. The product AL will be the tcrossprod
factor of a covariance matrix rather than a correlation matrix, unless a non-default normalization was requested when declaring the random-effect terms. User-defined random-effects models may also be heteroscedastic. In all these cases Corr
will by default return the correlation matrix, by applying cov2cor
to the tcrossproduct.
vcov
returns a matrix.
Corr
returns a list, for the different random effect terms. For each random-effect term with nontrivial correlation structure, the returned element is a matrix, returned in base matrix format or in some class from Matrix. Otherwise the it is an information message.
VarCorr
returns either NULL (if no variance to report, as for a poisson GLM) or a data frame with columns for the grouping factor, term, variance of random effect, standard deviation (the root of the variance), and optionally for correlation of random effect in random-coefficient terms. Information about the residual variance is optionally included as the last row(s) of the data frame, when relevant (gaussian- or Gamma-response models with single scalar parameter; beware the meaning of the residual variance parameter for Gamma-response models).
Some variance parameters may be removed from the VarCorr
output, with a message, such as the slope of the linear predictor describing the correlation model of an adjacency
term (see autoregressive
). The rare user of such parametrization should not consider this as a stable feature.
get_inits_from_fit
and get_ranPars
.
get_matrix(., which="beta_v_cov")
for the joint covariance matrix of estimates/predictions of fixed-effect coefficients and random effects; and get_matrix(., which="v_condcov")
for the covariance matrix of predictions of random effects given fixed effects (the variances corresponding to the condsd
reported in some lme4 output). Both of these computations refer to the random effects v as defined in random-effects
.
data("wafers") m1 <- HLfit(y ~ X1+X2+(1|batch), resid.model = ~ 1 ,data=wafers, method="ML") vcov(m1) # Example from VarCorr() documentation in 'nlme' package data("Orthodont",package = "nlme") sp1 <- fitme(distance ~ age+(age|Subject), data = Orthodont, method="REML") VarCorr(sp1)
data("wafers") m1 <- HLfit(y ~ X1+X2+(1|batch), resid.model = ~ 1 ,data=wafers, method="ML") vcov(m1) # Example from VarCorr() documentation in 'nlme' package data("Orthodont",package = "nlme") sp1 <- fitme(distance ~ age+(age|Subject), data = Orthodont, method="REML") VarCorr(sp1)
This (partially) documents the usage of the verbose
argument of the fitting functions, and more specifically of verbose["TRACE"]
values.
Default is TRACE=FALSE
(or 0) which is self-explanatory. TRACE=TRUE
(or 1) shows values of outer-estimated parameters (and possibly fixed values of parameters that would be outer-estimated), some cryptic progress bar, and the attained value of the likelihood objective function (but when there inner-estimated dispersion parameters, the output is more difficult to describe concisely). Other values have effect may change in later versions without notice see Details).
If the fitted model includes a residual-dispersion mdoel, some tracing output for the latter may be confusingly intermingled with tracing output of the mean-response model. The Details are valid only for the mean-response model.
0<TRACE
<1 only shows the cryptic progress bar.TRACE=2
will further show information about the progress of Levenberg-Marquardt-like steps for linear-predictor coefficients.TRACE=3
will further show information about the progress of distinct Levenberg-Marquardt-like steps random effects given fixed-effect coefficients.TRACE=4
will further show cryptic information about which matrix computations are requested.TRACE=5
will further report (through a call to the base trace
function) the str(.)
of the results of such matrix computations.TRACE=6
will further pause between iterations of the reweighted least-squares algorithm, allowing a browser session to be called.
This data set was reported and analyzed by Robinson et al. (2006) and reanalyzed by Lee et al. (2011). The data “deal with wafers in a single etching process in semiconductor manufacturing. Wafers vary through time since there are some variables that are not perfectly controllable in the etching process. For this reason, wafers produced on any given day (batch) may be different from those produced on another day (batch). To measure variation over batch, wafers are tested by choosing several days at random. In this data, resistivity is the response of interest. There are three variables, gas flow rate (x1), temperature (x2), and pressure (x3) and one random effect (batch or day).” (Lee et al 2011).
data("wafers")
data("wafers")
The data frame includes 198 observations on the following variables:
resistivity.
batch, indeed.
gas flow rate.
temperature.
pressure.
This data set was manually pasted from Table 3 of Lee et al. (2011). Transcription errors may have occurred.
Robinson TJ, Wulff SS, Montgomery DC, Khuri AI. 2006. Robust parameter design using generalized linear mixed models. Journal of Quality Technology 38: 38–65.
Lee, Y., Nelder, J.A., and Park, H. 2011. HGLMs for quality improvement. Applied Stochastic Models in Business and Industry 27, 315-328.
## see examples in the main Documentation page for the package.
## see examples in the main Documentation page for the package.
The data give the results of an unreplicated experiment for factors affecting welding quality conducted by the National Railway Corporation of Japan (Taguchi and Wu, 1980, cited in Smyth et al., 2001). It is a toy example for heteroscedastic models and is also suitable for illustrating fit of overparameterized models.
data("welding")
data("welding")
The data frame includes 16 observations on 10 variables:
response variable;
nine two-level factors.
The data were downloaded from http://www.statsci.org/data/general/welding.txt on 2014/08/19 and are consistent with those shown in table 5 of Bergman and Hynén (1997).
Bergman B, Hynén A (1997) Dispersion effects from
unreplicated designs in the series.
Technometrics, 39, 191–98.
Smyth GK, Huele AF, Verbyla AP (2001). Exact and approximate REML for heteroscedastic regression. Statistical Modelling 1, 161-175.
Taguchi G, Wu Y (1980) Introduction to off-line quality control. Nagoya, Japan: Central Japan Quality Control Association.
data("welding") ## toy example from Smyth et al. fitme(Strength ~ Drying + Material,resid.model = ~ Material+Preheating ,data=welding, method="REML") ## toy example of overparameterized model fitme(Strength ~ Rods+Thickness*Angle+(1|Rods),resid.model = ~ Rods+Thickness*Angle ,data=welding)
data("welding") ## toy example from Smyth et al. fitme(Strength ~ Drying + Material,resid.model = ~ Material+Preheating ,data=welding, method="REML") ## toy example of overparameterized model fitme(Strength ~ Rods+Thickness*Angle+(1|Rods),resid.model = ~ Rods+Thickness*Angle ,data=welding)
Translation of an example that may be found at
https://www.r-bloggers.com/2019/06/genotype-experiments-fitting-a-stability-variance-model-with-r/,
based on yield of eight durum wheat genotypes over seven years, following a randomised block design with three replicates.
A genotype-in-year random effect is used to quantify genotype-by-environment interactions.
In the first fit (constvar
), the variance of this random effect is constant over genotypes.
In the second fit (varvar
), different variances are fitted for the distinct genotypes, to assess the relative stability of yield of the different genotypes over environments. This second model can be fitted as a
constrained random-coefficient model, where the constraint describes a diagonal covariance matrix for the random coefficients.
This example uses the fact that the argument fixed=list(ranCoefs=<...>)
can be used to fit a
covariance matrix with an arbitrary set of constrained elements. Only elements
left as 'NA' (here the diagonal elements of the matrix) are fitted.
if (spaMM.getOption("example_maxtime")>1.5 && requireNamespace("agridat", quietly = TRUE)) { data("onofri.winterwheat", package="agridat") (constvar <- fitme( yield ~ gen + (1|year) + (1|block %in% year)+(1|gen %in% year), data=onofri.winterwheat, method="REML")) # Diagonal matrix of NA's, represented as vector for its lower triangle: ranCoefs_for_diag <- function(nlevels) { ## Conceptual version # diagmat <- matrix(NA, ncol=nlevels,nrow=nlevels) # diagmat[lower.tri(diagmat,diag=FALSE)] <- 0 # return(diagmat[lower.tri(diagmat,diag=TRUE)]) ## which amounts to: vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } (varvar <- fitme( yield ~ gen + (1|year) + (1|block %in% year)+(0+gen|gen %in% year), method="REML", data=onofri.winterwheat, fixed=list(ranCoefs=list("3"=ranCoefs_for_diag(8L))))) }
if (spaMM.getOption("example_maxtime")>1.5 && requireNamespace("agridat", quietly = TRUE)) { data("onofri.winterwheat", package="agridat") (constvar <- fitme( yield ~ gen + (1|year) + (1|block %in% year)+(1|gen %in% year), data=onofri.winterwheat, method="REML")) # Diagonal matrix of NA's, represented as vector for its lower triangle: ranCoefs_for_diag <- function(nlevels) { ## Conceptual version # diagmat <- matrix(NA, ncol=nlevels,nrow=nlevels) # diagmat[lower.tri(diagmat,diag=FALSE)] <- 0 # return(diagmat[lower.tri(diagmat,diag=TRUE)]) ## which amounts to: vec <- rep(0,nlevels*(nlevels+1L)/2L) vec[cumsum(c(1L,rev(seq(nlevels-1L)+1L)))] <- NA vec } (varvar <- fitme( yield ~ gen + (1|year) + (1|block %in% year)+(0+gen|gen %in% year), method="REML", data=onofri.winterwheat, fixed=list(ranCoefs=list("3"=ranCoefs_for_diag(8L))))) }
spaMM implements three interfaces for parallelisation. Depending on their arguments, either serial computation (default), a socket cluster (parallelisation default), or a fork cluster (available in linux and alike operating systems) can be used by all interfaces.
dopar
is called by default by its bootstrap procedures, and dofuture
has been developed as an alternative, whose use is controlled by spaMM.options(wrap_parallel="dofuture")
(versus the default, spaMM.options(wrap_parallel="dopar")
. combinepar
is the third and more recent interface; it is not a formally supported wrap_parallel
option because its additional functionalities are of no use in spaMM's bootstrap procedures.
dopar
is based on a patchwork of backends: for socket clusters, depending whether the doSNOW package is attached, foreach
or pbapply
is called (doSNOW allows more efficient load balancing than pbapply
); for fork clusters, parallel::mclapply
is used. This makes it impossible to ensure consistency of options accross computation environments, notably of enforcing the .combine
control of foreach
; and this makes it error-prone to ensure identical control of random number generators in all cases (although dopar
and combinepar
still aim to ensure the latter control).
By contrast, dofuture
is based only on the future and future.apply packages, in principle allowing a single syntax to control of random number generator across the different cases, hence repeatable results across them. This does not make a difference for bootstrap computations in spaMM as the bootstrap samples are never simulated in parallel: only refitting the models is performed in parallel, and fit results do not depend on random numbers. Further, the future-based code for socket clusters appears significantly slower than the one used by dopar
. For these reasons, the latter function is used by default by spaMM.
combinepar
is a third and more recent approach designed to address the other issue: it always uses foreach so that the .combine
control is consistently enforced. It uses future only when no alternative is available to produce a progress bar (namely, for socket clusters when doSNOW is not available).
X.GCA
and X.antisym
are functions which, when called in a model formula, stand for terms designed to represent the effet of symmetric interactions between pairs of individuals (order of individuals in the pair does not matter). antisym
likewise represents anti-symmetric interactions (the effect of reciprocal ordered pairs on the outcome are opposite, as in the so-called Bradley-Terry models). These constructs all account for multiple membership, i.e., the fact that the same individual may act as the first or the second individual among different pairs, or even within one pair if this makes sense).
The outcome of an interaction between a pair of agents is subject to a symmetric overall effect
when the effect “on” individual
(or viewed from the perspective of individual
) equals the effect on
:
. This may result from the additive effect of individual effects
and
:
. A
X.GCA
call represents such symmetric additive effects. Conversely, antisymmetry is characterized by and is represented by a
X.antisym
call. See the diallel
documentation for similar constructs for random effects, for additional comments on semantics (e.g. about “GCA”), and for further references.
If individual-level factors ID1 + ID2 were included in a formula for dyadic interactions, this would result in different coefficients being fitted for the same level in each factor. By contrast, the present constucts ensure that a single coefficient is fitted for the same-named levels of factors ID1 and ID2.
X.GCA(term, contr="contr.treatment", ...) X.antisym(term, contr="contr.treatment", ...)
X.GCA(term, contr="contr.treatment", ...) X.antisym(term, contr="contr.treatment", ...)
term |
an expression of the form <.>:<.> where each <.> represents a factor (or a variable that will automaticall be converted to a factor) present in the |
contr |
The |
... |
For programming purposes, not documented. |
The fixed-effect terms (GCA(Par1,Par2)
, etc) from the lmDiallel package (Onofri & Terzaroli, 2021), defined by functions returning design matrices for usage with stats:lm
, work in a formula for a spaMM fit, and users can define use such functions as templates for additional functions that will work similarly. However, not all post-fit functions will handle terms defined in this way well: checking robustness of predict
on small and permuted newdata
, as shown in the Examples, is a good test. Such problems happen because the formula-handling machinery of R handles terms represented by either a matrix or a factor, while both the model matrix and the factor information used to construct dyadic-interaction terms are needed to correctly predict, with new data, from models including such terms.
The presently designed functions are defined to solve this issue. By using such functions as template, users can define additional functions with the same return format (as further explained in the documented source code of X.antisym
), which will allow them to perform correct predictions from fitted models.
The functions return design matrices with additional class "factorS"
and attributes "call"
and "spec_levs"
.
Onofri A., Terzaroli N. (2021) lmDiallel: Linear fixed effects models for diallel crosses. Version 0.9.4. https://cran.r-project.org/package=lmDiallel.
#### Simulate dyadic data set.seed(123) nind <- 10 # Beware data grow as O(nind^2) x <- runif(nind^2) id12 <- expand.grid(id1=seq(nind),id2=seq(nind)) id1 <- id12$id1 id2 <- id12$id2 u <- rnorm(nind,mean = 0, sd=0.5) ## additive individual effects: y <- 0.1 + 1*x + u[id1] + u[id2] + rnorm(nind^2,sd=0.2) ## anti-smmetric individual effects: t <- 0.1 + 1*x + u[id1] - u[id2] + rnorm(nind^2,sd=0.2) dyaddf <- data.frame(x=x, y=y, t=t, id1=id1,id2=id2) # : note that this contains two rows per dyad, which avoids identifiability issues. # Enforce that interactions are between distinct individuals (not essential for the fit): dyaddf <- dyaddf[- seq.int(1L,nind^2,nind+1L),] # Fits: (addfit <- fitme(y ~x +X.GCA(id1:id2), data=dyaddf)) (antifit <- fitme(t ~x +X.antisym(id1:id2), data=dyaddf)) if (FALSE) { #### check of correct handling by predict(): # First scramble the data so that input factors are in no particular order set.seed(123) dyaddf <- dyaddf[sample(nrow(dyaddf)),] addfiti <- fitme(y ~x +X.GCA(id1:id2), data=dyaddf) foo <- rev(2:4) p1 <- predict(addfiti)[foo] p2 <- predict(addfiti, newdata=dyaddf[foo,]) diff(range(p1-p2))<1e-10 # must be TRUE }
#### Simulate dyadic data set.seed(123) nind <- 10 # Beware data grow as O(nind^2) x <- runif(nind^2) id12 <- expand.grid(id1=seq(nind),id2=seq(nind)) id1 <- id12$id1 id2 <- id12$id2 u <- rnorm(nind,mean = 0, sd=0.5) ## additive individual effects: y <- 0.1 + 1*x + u[id1] + u[id2] + rnorm(nind^2,sd=0.2) ## anti-smmetric individual effects: t <- 0.1 + 1*x + u[id1] - u[id2] + rnorm(nind^2,sd=0.2) dyaddf <- data.frame(x=x, y=y, t=t, id1=id1,id2=id2) # : note that this contains two rows per dyad, which avoids identifiability issues. # Enforce that interactions are between distinct individuals (not essential for the fit): dyaddf <- dyaddf[- seq.int(1L,nind^2,nind+1L),] # Fits: (addfit <- fitme(y ~x +X.GCA(id1:id2), data=dyaddf)) (antifit <- fitme(t ~x +X.antisym(id1:id2), data=dyaddf)) if (FALSE) { #### check of correct handling by predict(): # First scramble the data so that input factors are in no particular order set.seed(123) dyaddf <- dyaddf[sample(nrow(dyaddf)),] addfiti <- fitme(y ~x +X.GCA(id1:id2), data=dyaddf) foo <- rev(2:4) p1 <- predict(addfiti)[foo] p2 <- predict(addfiti, newdata=dyaddf[foo,]) diff(range(p1-p2))<1e-10 # must be TRUE }
A ZAXlist
object is a representation of the “ZAL
” matrix as an S4 class holding a list of descriptors of each ZAL
block for each random effect.
A Kronfacto
object is a representation of a Kronecker product as an S4 class holding its factors. Methods defined for this class may avoid the computation of the Kronecker product as an actual matrix of large dimensions.
This documentation is for development purposes and may be incomplete. The objects and methods are not part of the programming interface and are subject to modification without notice.
# new("ZAXlist", LIST=.) # new("Kronfacto", BLOB=.)
# new("ZAXlist", LIST=.) # new("Kronfacto", BLOB=.)
LIST
:A list whose each element is a ZAL block represented as either a (M|m)atrix
,
or a list with two elements (and additional class ZA_QCHM
): ZA
, and the Cholesky
factor Q_CHMfactor
of the precision matrix (L=solve(Q_CHMfactor,system="Lt")
).
BLOB
:An environment holding lhs
and rhs
, the factors of the Kronecker product, and other objects initialized as promises. See the source code of the non-exported .def_Kranfacto
constructor for further information.