R/getXB.R

In momentuHMM: Maximum Likelihood Analysis of Animal Movement Behavior Using Multivariate Hidden Markov Models

getXB<-function(DM,nbObs,wpar,DMind,circularAngleMean,consensus,nbStates,nc,meanind){
  nrw <- nrow(DM)
  ncl <- ncol(DM)
  if(isFALSE(circularAngleMean)){
    if(DMind){
      XB <- DM%*%wpar
    } else {
      XB <- XBloop_rcpp(DM, wpar, nbObs, nrw, ncl, FALSE, FALSE, 1:nrw-1, nc, nbStates)
    }
  } else {
    if(length(meanind)) {
      XB<-XBloop_rcpp(DM,wpar,nbObs,nrw,ncl,TRUE,consensus,meanind-1,nc,nbStates,as.numeric(circularAngleMean))
      if(consensus) {
        l_t <- matrix(1,nrow(XB),ncol(XB))
        l_t[nbStates+meanind,] <- XB[nbStates+meanind,]
      }
    } else {
      XB <- matrix(0,nrw,nbObs)
      if(consensus) l_t <- matrix(1,nrow(XB),ncol(XB))
    }
    cInd <- which(colSums(nc[nbStates+1:nbStates,,drop=FALSE])>0)
    wpar2 <- wpar[cInd - (ncl-length(wpar))]
    XB[nbStates+1:nbStates,]<-XBloop_rcpp(DM[nbStates+1:nbStates,cInd,drop=FALSE],wpar2,nbObs,nbStates,length(wpar2),FALSE,FALSE,1:nbStates-1,nc[nbStates+1:nbStates,cInd,drop=FALSE],nbStates)
  }
  if(consensus){
    return(list(XB=XB,l_t=l_t))
  } else return(XB)
}