If you need an accessible version of this item please contact JSTOR User Support

Markov Chain Monte Carlo Convergence Diagnostics: A Comparative Review

Mary Kathryn Cowles and Bradley P. Carlin
Journal of the American Statistical Association
Vol. 91, No. 434 (Jun., 1996), pp. 883-904
DOI: 10.2307/2291683
Stable URL: http://www.jstor.org/stable/2291683
Page Count: 22
  • Download PDF
  • Cite this Item

You are not currently logged in.

Access your personal account or get JSTOR access through your library or other institution:

login

Log in to your personal account or through your institution.

If you need an accessible version of this item please contact JSTOR User Support
Markov Chain Monte Carlo Convergence Diagnostics: A Comparative Review
Preview not available

Abstract

A critical issue for users of Markov chain Monte Carlo (MCMC) methods in applications is how to determine when it is safe to stop sampling and use the samples to estimate characteristics of the distribution of interest. Research into methods of computing theoretical convergence bounds holds promise for the future but to date has yielded relatively little of practical use in applied work. Consequently, most MCMC users address the convergence problem by applying diagnostic tools to the output produced by running their samplers. After giving a brief overview of the area, we provide an expository review of 13 convergence diagnostics, describing the theoretical basis and practical implementation of each. We then compare their performance in two simple models and conclude that all of the methods can fail to detect the sorts of convergence failure that they were designed to identify. We thus recommend a combination of strategies aimed at evaluating and accelerating MCMC sampler convergence, including applying diagnostic procedures to a small number of parallel chains, monitoring autocorrelations and cross-correlations, and modifying parametrizations or sampling algorithms appropriately. We emphasize, however, that it is not possible to say with certainty that a finite sample from an MCMC algorithm is representative of an underlying stationary distribution.

Page Thumbnails

  • Thumbnail: Page 
883
    883
  • Thumbnail: Page 
884
    884
  • Thumbnail: Page 
885
    885
  • Thumbnail: Page 
886
    886
  • Thumbnail: Page 
887
    887
  • Thumbnail: Page 
888
    888
  • Thumbnail: Page 
889
    889
  • Thumbnail: Page 
890
    890
  • Thumbnail: Page 
891
    891
  • Thumbnail: Page 
892
    892
  • Thumbnail: Page 
893
    893
  • Thumbnail: Page 
894
    894
  • Thumbnail: Page 
895
    895
  • Thumbnail: Page 
896
    896
  • Thumbnail: Page 
897
    897
  • Thumbnail: Page 
898
    898
  • Thumbnail: Page 
899
    899
  • Thumbnail: Page 
900
    900
  • Thumbnail: Page 
901
    901
  • Thumbnail: Page 
902
    902
  • Thumbnail: Page 
903
    903
  • Thumbnail: Page 
904
    904