Accession Number : ADA308874
Title : Easily Verifiable Conditions for the Convergence of the Markov Chain Monte Carlo Method.
Descriptive Note : Technical rept.,
Corporate Author : FLORIDA STATE UNIV TALLAHASSEE
Personal Author(s) : Sethuraman, Jayaram
PDF Url : ADA308874
Report Date : DEC 1995
Pagination or Media Count : 12
Abstract : The Markov Chain Chain Monte Carlo (MCMC) method, which is a special case of the Gibbs sampler, is a very powerful method to simulate from complicated distributions arising in many contexts, including image analysis, computational Bayesian analysis, and so on. Existing results that ensure that this method will converge involve conditions which are difficult to verify in practice, and most practitioners, convinced that their particular problem will not be pathological and give up verifying altogether. This paper gives a new set of sufficient conditions which are easy to verify in most applications.
Descriptors : *MONTE CARLO METHOD, *MARKOV PROCESSES, MAXIMUM LIKELIHOOD ESTIMATION, PROBABILITY DISTRIBUTION FUNCTIONS, RANDOM VARIABLES, STATISTICAL DATA, PROBABILITY DENSITY FUNCTIONS, CONVERGENCE, BAYES THEOREM, POPULATION(MATHEMATICS).
Subject Categories : Statistics and Probability
Distribution Statement : APPROVED FOR PUBLIC RELEASE