Hi Martin,
Can I use JAGS to approximate the Marginal Likelihood of a hierarchical model?
I am aware that I can approximate DIC in JAGS for comparing hierarchical models, but I could not find how to get the Bayes Factor!
Would you please advise me ?
Thank you,
Garo.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Hi Martin,
Can I use JAGS to approximate the Marginal Likelihood of a hierarchical model?
I am aware that I can approximate DIC in JAGS for comparing hierarchical models, but I could not find how to get the Bayes Factor!
Would you please advise me ?
Thank you,
Garo.
It is incredibly hard to get Bayes factors from MCMC. See this review on the various special techniques that may be required.
http://www.stat.columbia.edu/~gelman/research/published/path2.pdf
Thank you Martin for your reply. Indeed, I thought of the power posterior method (bridge sampling). I will give a try.
cheers,
Garo.