The increasing number of R-oriented Bayesian computational tools such as
MCMCpack,
MCMCglmm,
DPpackage,
R-INLA,
spBayes, have made BUGS less and less crucial for day to day Bayesian computation. Honestly, I cannot figure out a single analysis that BUGS can do but at least one of the above mentioned packages cannot.
From a quick look at the web site,
R-INLA seems to be able to carry out bayesian analysis without the actual MCMC sampling. This makes it possible to quickly try various different model specifications without hours of waiting time. This package looks very interesting and definitely deserves more attention.
http://sgsong.blogspot.com/search/label/Bayesian
没有评论:
发表评论