R/nextF.R

In RSGHB: Functions for Hierarchical Bayesian Estimation: A Flexible Approach

nextF = function(p, f, b, env) {
  
  # applying an MH algo here  
  fnew <- f + sqrt(env$rhoF) * rnorm(env$gFIV, 0, 1)
  pnew <- env$likelihood(fnew, b, env)
  
  r <- sum(log(pnew) - log(p))
  ind <- (log(runif(1)) <= r)
  
  pnew <- pnew * ind + p * (1 - ind)
  fnew <- as.vector(t(fnew %*% t(ind)) + t(f %*% t(1 - ind)))
  
  return(list(fnew, pnew))
}