R Package MCMC and Potts

R Package MCMC

Package

At CRAN.

This web page is about an R package for doing simple, but general MCMC. It does random-walk Metropolis for an arbitrary continuous distribution on R^d specifed by an unnormalized density computed by a user-supplied R function.

For more info see

• the package vignette,
• the design document, or
• or the R help pages for the functions
  • metrop which simulates the distribution of any continuous random vector via a normal random walk Metropolis algorithm.
  • temper which simultaneously simulates several distributions of continuous random vectors using parallel or serial simulated tempering.
  • initseq which does variance estimation using the initial sequence estimators of Geyer (1992).
  • olbm which does variance estimation using the method of overlapping batch means.

The source code for the library is

Talk

Not just about this package, but made to go with it are the slides for the talk given at the Institute for Mathematics and its Application Sep 16, 2003

R Package Potts

In a separate package potts, also at CRAN, is an implementation of the Swendsen-Wang algorithm for Potts models.