PS. sorry, obviously that should have been rpois(1...

2008-03-24T19:31:00.000-07:00

PS. sorry, obviously that should have been rpois(1, 500*sample(lambda, 1))

Hi Jim,For the posterior predictive distribution y...

2008-03-24T19:28:00.000-07:00

Hi Jim,
For the posterior predictive distribution you have:

ys=rpois(1000,500*lambda)

where lambda was previously calculated as:

lambda=rgamma(1000,shape=y,rate=t)

I'm a bit unsure about this. Isn't the rpois call only using the first element of the lambda vector ? I would have thought to do something like this:
ys <- numeric(1000)
for ( i in 1:1000 ) ys[i] <- rpois(1, sample(lambda,1))

Am I missing something ?

Comments on Introduction to Bayesian Thinking: Inference for a Poisson Rate

PS. sorry, obviously that should have been rpois(1...

Hi Jim,For the posterior predictive distribution y...