Harris recurrence of Metropolis-within-Gibbs and trans-dimensional Markov chains

DOI10.1214/105051606000000510zbMATH Open1121.60076arXivmath/0702412OpenAlexW1988251813MaRDI QIDQ997421FDOQ997421

Authors: Gareth O. Roberts, Jeffrey S. Rosenthal 

Publication date: 6 August 2007

Published in: The Annals of Applied Probability (Search for Journal in Brave)

Abstract: A $phi$-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.

Full work available at URL: https://arxiv.org/abs/math/0702412

Recommendations

• On the convergence of the Metropolis-Hastings Markov chains
• Rates of convergence of the Hastings and Metropolis algorithms
• A new proof of convergence of MCMC via the ergodic theorem
• General state space Markov chains and MCMC algorithms
• Markov Chain Monte Carlo Algorithms: Theory and Practice

zbMATH Keywords

Markov chain Monte CarloMetropolis algorithmHarris recurrencephi-irreducibilitytrans-dimensional Markov chains

Mathematics Subject Classification ID

Computational methods in Markov chains (60J22) Bayesian inference (62F15) Monte Carlo methods (65C05) Discrete-time Markov processes on general state spaces (60J05)

Cites Work

• Reversible jump Markov chain Monte Carlo computation and Bayesian model determination
• Markov chains for exploring posterior distributions. (With discussion)
• Sampling-Based Approaches to Calculating Marginal Densities
• Title not available (Why is that?)
• Monte Carlo sampling methods using Markov chains and their applications
• Title not available (Why is that?)
• A note on Metropolis-Hastings kernels for general state spaces
• Markov chains and stochastic stability
• General state space Markov chains and MCMC algorithms
• General Irreducible Markov Chains and Non-Negative Operators
• Efficient Construction of Reversible Jump Markov Chain Monte Carlo Proposal Distributions
• Minorization Conditions and Convergence Rates for Markov Chain Monte Carlo
• Bounds on regeneration times and convergence rates for Markov chains
• Honest exploration of intractable probability distributions via Markov chain Monte Carlo.
• Convergence Properties of Perturbed Markov Chains
• Quantitative convergence rates of Markov chains: A simple account
• Title not available (Why is that?)
• Title not available (Why is that?)
• Convergence Rates for Markov Chains
• Title not available (Why is that?)
• Title not available (Why is that?)

Cited In (19)

• Objective Bayesian analysis of degradation model with respect to a Wiener process
• Model choice using reversible jump Markov chain Monte Carlo
• Unbiased inference for discretely observed hidden Markov model diffusions
• Bayesian estimation in Kibble's bivariate gamma distribution
• Micro-local analysis for the Metropolis algorithm
• Bayesian Dirichlet mixture model for multivariate extremes: a re-parametrization
• Analysis of two-component Gibbs samplers using the theory of two projections
• A Comprehensive Bayesian Framework for Envelope Models
• Monte Carlo methods for light propagation in biological tissues
• A Bayesian Approach to Envelope Quantile Regression
• Importance sampling correction versus standard averages of reversible MCMCs in terms of the asymptotic variance
• Ergodicity of combocontinuous adaptive MCMC algorithms
• Ergodicity of Markov processes via nonstandard analysis
• Time series with infinite-order partial copula dependence
• Wasserstein-based methods for convergence complexity analysis of MCMC with applications
• A quasi-Bayesian perspective to online clustering
• Parallel tempering with equi-energy moves
• Typical representatives of free homotopy classes in multi-punctured plane
• Maximin design on non hypercube domains and kernel interpolation

This page was built for publication: Harris recurrence of Metropolis-within-Gibbs and trans-dimensional Markov chains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q997421)