fcondYZXA: Conditional posterior distribution of the bivariate latents...

In BNPdensity: Ferguson-Klass Type Algorithm for Posterior Normalized Random Measures

Description

This function simulates form the conditional posterior distribution of the latents (Y,Z).

Usage

fcondYZXA(x, distr, Tauy, Tauz, J)

Details

For internal use.

Examples

## The function is currently defined as
function (x, distr = 1, Tauy, Tauz, J) 
{
    K <- matrix(NA, nrow = length(Tauy), ncol = length(x))
    for (i in seq(Tauy)) {
        K[i, ] <- dk(x, distr = distr, mu = Tauy[i], sigma = Tauz[i]) * 
            J[i]
    }
    if (any(is.na(K))) 
        print(K, Tauy, Tauz, J)
    pK <- prop.table(K, margin = 2)
    j <- apply(pK, 2, function(x) sample(length(Tauy), size = 1, 
        prob = x))
    return(matrix(c(y = Tauy[j], z = Tauz[j]), nrow = length(x), 
        ncol = 2))
  }

BNPdensity documentation built on May 29, 2017, 9:33 p.m.

Conditional posterior distribution of the latents Y in the...

BNPdensity index

Conditional posterior distribution of the bivariate latents...