# gs4: Resampling Ystar function In BNPdensity: Ferguson-Klass Type Algorithm for Posterior Normalized Random Measures

## Description

This funciton resamples the distinct Ystar in the semiparametric model.

## Usage

 `1` ```gs4(ystar, x, idx, distr.k, sigma.k, distr.p0, mu.p0, sigma.p0) ```

## Details

For internal use.

## Examples

 ``` 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20``` ```## The function is currently defined as function (ystar, x, idx, distr.k, sigma.k, distr.p0, mu.p0, sigma.p0) { r <- length(ystar) nstar <- as.numeric(table(idx)) for (j in seq(r)) { id <- which(!is.na(match(idx, j))) xj <- x[id] xbar <- sum(xj)/nstar[j] y2star <- rk(1, distr = distr.k, mu = xbar, sigma = sigma.k/sqrt(nstar[j])) f.ratio <- rfyzstar(y2star, ystar[j], xj, distr = distr.k, sigma = sigma.k, distr.p0 = distr.p0, mu.p0 = mu.p0, sigma.p0 = sigma.p0) k.ratio <- dk(ystar[j], distr = distr.k, mu = xbar, sigma = sigma.k/sqrt(nstar[j]))/dk(y2star, distr = distr.k, mu = xbar, sigma = sigma.k/sqrt(nstar[j])) q2 <- min(1, f.ratio * k.ratio) ystar[j] <- ifelse(runif(1) <= q2, y2star, ystar[j]) } return(ystar) } ```

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