MCMC for probabilistic graphical models in R

In fact, this section could be the title of book. As a matter of fact, there are several books entirely devoted to this specific topic. Research in this field is extremely active, with many new algorithms coming every year.

There are numerous packages implementing MCMC algorithms for different types of algorithms. There are also more generic frameworks such as the famous BUGS (and its open source implementation OpenBUGS) and a new, even more powerful framework called Stan. Historically, BUGS was the first framework to popularize MCMC inference in Bayesian statistics and literally led to a revolution in this field, as everyone could benefit from Bayesian statistics right out of the box.

Making an introduction to each of them and showing all the possibilities of MCMC for a few specific graphical models would require another book as big as this one. In this section, we will therefore focus on a programming environment available in R. In fact, it works...