R/lposttau.R

In jcgaasch/LaRA: Latent Regression Analysis

lposttau <- function(
  Tau,
  Yj,
  Qj,
  alpha,
  beta,
  Theta
){

  Kappa <- c(-1e+05, 0, cumsum(exp(Tau)), 1e+05)
  lprior <- dmvnorm(Tau, rep(0, Qj), 100*diag(Qj), TRUE)
  llik <- sum(log(pnorm(alpha*Theta - (beta + Kappa[Yj + 1])) -
    pnorm(alpha*Theta - (beta + Kappa[Yj + 2]))))
  lpost <- lprior + llik
  return(-lpost)

}