PROC MCMC and informative priors
proc mcmc data=ONFIT seed=27500 nbi=5000 ntu=5000 nmc=200000
thin=10;
parms beta0 beta1 beta2 beta3;
prior beta1 ~ normal(0.01, var=0.01);
prior beta0 beta2 beta3 ~ normal(0, var=10000);
p=logistic(beta0+beta1*CDD+beta2*RAIN+beta3*WW);
model ONFIT ~ binomial(N,p);
title "Bayesian Analysis using MCMC - informative priors";
run;
Suggested exercises after completing case study #3:
(a) Compare results from case study
#2 and case study #4. In particular, readers should look at the plots (MCMC chain, autocorrelation and density plots) and the Posterior summaries. Similar to when cases
#1 and
#3 were compared, results from case
#2 and #4 should look very similar.
