R/effecush.R

In CUB: A Class of Mixture Models for Ordinal Data

#' @title Auxiliary function for the log-likelihood estimation of CUSH models with covariates
#' @description Compute the opposite of the loglikelihood function for CUSH models
#' with covariates to explain the shelter effect.
#' @aliases effecush
#' @usage effecush(paravec, esternocush, shelter, m)
#' @param paravec Vector of the initial parameters estimates
#' @param esternocush Matrix binding together the vector of ordinal data and the matrix XX of explanatory
#' variables whose first column is a column of ones needed to consider an intercept term
#' @param shelter Category corresponding to the shelter choice
#' @param m Number of ordinal categories
#' @keywords internal 
#' @details It is called as an argument for "optim" within CUSH function (when no covariate is included)
#'  as the function to minimize.


effecush <-function(paravec,esternocush,shelter,m){
  ordinal<-esternocush[,1]
  ncovar<-ncol(esternocush)
  X<-esternocush[,3:ncovar]
  return(-loglikcushcov(m,ordinal,X,paravec,shelter))
}