Annals of Applied Probability
- Ann. Appl. Probab.
- Volume 16, Number 4 (2006), 2123-2139.
Harris recurrence of Metropolis-within-Gibbs and trans-dimensional Markov chains
A ϕ-irreducible and aperiodic Markov chain with stationary probability distribution will converge to its stationary distribution from almost all starting points. The property of Harris recurrence allows us to replace “almost all” by “all,” which is potentially important when running Markov chain Monte Carlo algorithms. Full-dimensional Metropolis–Hastings algorithms are known to be Harris recurrent. In this paper, we consider conditions under which Metropolis-within-Gibbs and trans-dimensional Markov chains are or are not Harris recurrent. We present a simple but natural two-dimensional counter-example showing how Harris recurrence can fail, and also a variety of positive results which guarantee Harris recurrence. We also present some open problems. We close with a discussion of the practical implications for MCMC algorithms.
Ann. Appl. Probab., Volume 16, Number 4 (2006), 2123-2139.
First available in Project Euclid: 17 January 2007
Permanent link to this document
Digital Object Identifier
Mathematical Reviews number (MathSciNet)
Zentralblatt MATH identifier
Primary: 60J05: Discrete-time Markov processes on general state spaces
Secondary: 65C05: Monte Carlo methods 60J22: Computational methods in Markov chains [See also 65C40] 62F15: Bayesian inference
Roberts, Gareth O.; Rosenthal, Jeffrey S. Harris recurrence of Metropolis-within-Gibbs and trans-dimensional Markov chains. Ann. Appl. Probab. 16 (2006), no. 4, 2123--2139. doi:10.1214/105051606000000510. https://projecteuclid.org/euclid.aoap/1169065219