Bounding the convergence time of the Gibbs sampler in Bayesian image restoration

Author: Gibbs A.L.

Source: Biometrika, Volume 87, Number 4, December 2000 , pp. 749-766(18)

Purchase options

The full text electronic article is available for purchase. You will be able to download the full text electronic article after payment.

$33.35 plus tax Refund Policy

More like this?

In this: publication
By this: publisher
In this Subject: Biology , Public Health
By this author: Gibbs A.L.

Content Key:

- Free

- New

- Open Access

- Subscribed

- Free Trial

Abstract:

This paper shows how coupling methodology can be used to give precise, a priori bounds on the convergence time of Markov chain Monte Carlo algorithms for which a partial order exists on the state space which is preserved by the Markov chain transitions. This methodology is applied to give a bound on the convergence time of the random scan Gibbs sampler used in the Bayesian restoration of an image of N pixels. For our algorithm, in which only one pixel is updated at each iteration, the bound is a constant times N². The proportionality constant is given and is easily calculated. These bounds also give an indication of the running time of coupling from the past algorithms.

Keywords: Bayesian image restoration; Convergence; Coupling

Language: English

Document Type: Original article

Affiliations: Department of Mathematics and Statistics, York University, Toronto, Ontario, Canada M3J 1P3 agibbs@mathstat.yorku.ca: