mkGeneralGlmerDevfun: Make a general (Laplace) glmer deviance function

In stevencarlislewalker/lme4ord: Generalized Linear Mixed Models with Flexible (Co)Variance Structures

Description

Make a general (Laplace) glmer deviance function

Usage

mkGeneralGlmerDevfun(y, X, Zt, Lambdat, weights = NULL, offset = NULL,
  etastart = NULL, mustart = NULL, devfunEnv, initPars, parInds,
  mapToCovFact = NULL, mapToModMat = NULL, mapToWeights = NULL,
  penFixef = NULL, penCovar = NULL, penLoads = NULL, family,
  tolPwrss = 1e-06, maxit = 30, verbose = 0L, Lind = NULL)

Arguments

y	response vector
X	fixed effects model matrix
Zt	transposed random effects model matrix
Lambdat	relative covariance factor
weights	prior weights
offset	known additive offsets
etastart,mustart	starting values for linear predictor (etastart) or mean (mustart)
devfunEnv	environment of the returned deviance function
initPars	initial values for the parameter vector
parInds	a named list of vectors for identifying each of the three types of parameters in initPars, one vector for each of the three types of parameters: list(covar = ., fixef = ., loads = .) where the .'s are vectors of indices
mapToCovFact	function taking the covar parameters and returning the values of the non-zero elements of Lambdat (i.e. the x slot of Lambdat)
mapToModMat	function taking the loads parameters and returning the values of the non-zero elements of Zt (i.e. the x slot of Zt)
mapToWeights	function taking the weigh parameters
penFixef,penCovar,penLoads	optional functions of each parameter type, returning a scalar penalty for the deviance function
family	family
tolPwrss	tolerance for penalized weighted residual sum of squares
verbose	verbose
maxit	maximum number of PIRLS iterations
Lind	optional indices mapping covariance parameters to the non-zero elements of the relative covariance factor

Value

a deviance function with an environment

stevencarlislewalker/lme4ord documentation built on May 30, 2019, 4:43 p.m.