The R Companion to MCSM (6)

The chapter on the Metropolis algorithm is now completed and we are thus tantalisingly close to the end! (The end of the complete draft…) The completed chapters are

1. Introduction to R programming
2. Random variable generation
3. Monte Carlo methods
4. Controlling and accelerating convergence
5. Monte Carlo optimization
6. Metropolis-Hastings algorithms
7. [=8] Convergence monitoring for MCMC algorithms

We have now reached 226 pages, 72K words (whatever that means using wc on a pdf file), and 69 figures… There still seems to be a possibility that the final chapter on Gibbs sampling could be ready by the end of the month, if both George and I make rush for it. There is no major surprise in the current chapter, with independent Metropolis and random walk Metropolis algorithms being the central heroes, but there still was work to be done, in terms of coding new examples. We also included a nice Bayesian model choice illustration inspired from Bayesian Core where the Metropolis-Hastings algorithm only moves over model indices, the parameters being analytically integrated out. Once the first draft is over, the next stepstone will be the design of the corresponding package, collecting and classifying our R codes and few datasets…

PS—The title preferred by those who voted on the poll seems to be Monte Carlo Methods with R.