Purpose

---

  This class defines an interface for running Markov Chain Monte
  Carlo simulations. The code is based on the opening chapters of
  "Markov Chain Monte Carlo in Practice" by Gilks et al. The code in
  this file sets up the logic for running an MCMC simulation using
  the Metropolis Hastings algorithm. To actually use these functions,
  a subclass must be defined, and the selectTrial and
  logTargetDistribution methods must be overridden. Each of these
  methods is problem-specific, and cannot be coded in advance.

Category

---

  statistics

Modification History

---

  September 2009: Written by Chris Beaumont
  October 2 2009: Modified run procedure to avoid unnecessary calls
  to logTargetDistribution
  October 3 2009: Added THIN keyword and state variable