R/nextDind.R

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

nextDind = function(a, b, env) {
  
  d <- matrix(0, env$gNIV, env$gNIV)
  b <- matrix(t(a), nrow = env$gNP, ncol = env$gNIV, byrow = TRUE) - b
    
  for (k in 1:env$gNIV) {
    t <- 1 + t(b[, k]) %*% b[, k]
    s <- sqrt(1/t)[1, 1] * matrix(rnorm(env$gNP + 1), nrow = env$gNP + 1, ncol = 1)
    d[k, k] <- solve(t(s) %*% s)
  }
  
  return(d)
}