hsemfit {hsem}

Fitting Hierarchical Structural Equation Models using h-likelihood Approach

Description

The hsemfit is used to fit a hierarchical structural equation models (HSEMs) allowing different models for multivariate reponse variables. A variety of distributions and link functions for both response and the random effects are allowed. To call the fitting function hsemfit, models for the mean and dispersion must be specified by hsemmodleing object preferably created by calling the hsemmodeling function.

Usage

hsemfit(RespDist = "gaussian", BinomialDen = NULL, 
DataMain, MeanModel,DispersionModel = NULL, 
PhiFix = NULL, LamFix = NULL, structure = "correlated", 
mord = 0, dord = 1, convergence = 1e-05, 
Init_Corr = NULL, EstimateCorrelations = TRUE)

Arguments

RespDist

The distribution of the response is set by the option RespDist. The user can set it to: "gaussian" (default), "binomial", "poisson", or "gamma".

BinomialDen

When RespDist="binomial", one should use the option BinomialDen to specify the denominator for the binomial distribution. This should be "NULL" (default) or a numeric vector of length equal to the length of DataMain. When specified as BinomialDen=NULL and RespDist="binomial", the denominator is 1.

DataMain

The option DataMain determines the data frame to be used (non-optional).

MeanModel

For the mean model, this option requries DGHLMMODELING object which should specified by the option Model="mean".

DispersionModel

For the overdispersion model, this option requries DGHLMMODELING object which should be specified by the option Model="dispersion".

PhiFix

The option for overdispersion parameters (phi) to be estimated or maintaned constant. Specifying defaults such as PhiFix =NULL implies that phi is to be estimated. If not, phi is fixed at a value specified by PhiFix.

LamFix

The option for random-effect variance (lambda) to be estimated or maintaned constant. Specifying defaults such as LamFix =NULL implies that lambda is to be estimated. If not, lambda is fixed at a value specified by LamFix.

structure

The option structure determines structure of random effects. When structure="correlated" (or "shared"), correlated (or shared) random-effects model is specified.

mord

The option mord specifies the order of Laplace approximation to the marginal likelihood for fitting the mean parameters. The choice is either 0 or 1 (default).

dord

The option dord specifies the order of adjusted profile likelihood for fitting the dispersion parameters. The choice is either 1 (default) or 2.

convergence

Setting this option determines the criterion for convergence, which is computed as the absolute difference between the values of all the estimated parameters in the previous and current iterations. The default criterion is 1e-06.

Init_Corr

Setting initial values of correlation (or shared parameters) between random effects

EstimateCorrelations

Correlation are estimated or fixed when EstimateCorrelations=TRUE (default) or EstimateCorrelations=FALSE.

Value

res

the output class resulted from jointfit_correlated.

Examples

data(ml2)

MM1<-hsemmodeling(Model="mean",Link="identity",
   LinPred=urge~urge1+dep1+(1|id)+(urge1|id)+(dep1|id),
   RandDist=c("gaussian","gaussian","gaussian"))
DM1<-hsemmodeling(Model="dispersion",Link = "log", 
   LinPred=phi~(1|id),RandDist=c("gaussian"))
MM2<-hsemmodeling(Model="mean",Link="identity",
   LinPred=dep~urge1+dep1+(1|id)+(urge1|id)+(dep1|id),
   RandDist=c("gaussian","gaussian","gaussian"))
DM2<-hsemmodeling(Model="dispersion",Link = "log", 
   LinPred=phi~(1|id),RandDist=c("gaussian"))
res<-hsemfit(RespDist=c("gaussian","gaussian"),DataMain=list(ml2,ml2),
   structure="independent",MeanModel=list(MM1,MM2),
   DispersionModel=list(DM1,DM2))