[R] Is there a way to not use an explicit loop?

[Dan Davison]

Both shape parameters of rbeta can be vectors; for

x <- rbeta(n, shape1, shape2)

x[i] ~ Beta(shape1[i], shape2[i])

so

bbsim <- function(m=1000, num.post.draws=1e4, size.a=100, prob.a=.27, prior.count=1) {
    data.count <- rbinom(m, size.a, prob.a)
    shape1 <- rep(prior.count + data.count, each=num.post.draws)
    shape2 <- rep(prior.count + size.a - data.count, each=num.post.draws)
    matrix(rbeta(m * num.post.draws, shape1, shape2), num.post.draws, m)
}

Then you can do

beta.draws <- bbsim()
means <- apply(beta.draws, 2, mean)
medians <- apply(beta.draws, 2, median)
etc

Dan

On Wed, Sep 17, 2008 at 11:56:36AM -0700, Juancarlos Laguardia wrote:
> I have a problem in where i generate m independent draws from a binomial 
> distribution,
> say
>
> draw1 = rbinom( m , size.a, prob.a )
>
>
> then I need to use each draw to generate a beta distribution.  So, like 
> using a beta prior, binomial likelihood, and obtain beta posterior, m 
> many times.  I have not found out a way to vectorize draws from a beta 
> distribution, so I have an explicit for loop within my code
>
>
>
> for( i in 1: m ) {
>
> beta.post = rbeta( 10000, draw1[i] + prior.constant  ,  prior.constant + 
> size.a  - draw1[i] )
>
> beta.post.mean[i] = mean(beta.post)
> beta.post.median[i] = median(beta.post)
>
> etc.. for other info
>
> }
>
> Is there a way to vectorize draws from an beta distribution?
>
> UC Slug
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.

-- 
http://www.stats.ox.ac.uk/~davison