Translator Disclaimer
December 2008 EM versus Markov chain Monte Carlo for estimation of hidden Markov models: acomputational perspective
Tobias Rydén
Bayesian Anal. 3(4): 659-688 (December 2008). DOI: 10.1214/08-BA326

Abstract

Hidden Markov models (HMMs) and related models have become standard in statistics during the last 15--20 years, with applications in diverse areas like speech and other statistical signal processing, hydrology, financial statistics and econometrics, bioinformatics etc. Inference in HMMs is traditionally often carried out using the EM algorithm, but examples of Bayesian estimation, in general implemented through Markov chain Monte Carlo (MCMC) sampling are also frequent in the HMM literature. The purpose of this paper is to compare the EM and MCMC approaches in three cases of different complexity; the examples include model order selection, continuous-time HMMs and variants of HMMs in which the observed data depends on many hidden variables in an overlapping fashion. All these examples in some way or another originate from real-data applications. Neither EM nor MCMC analysis of HMMs is a black-box methodology without need for user-interaction, and we will illustrate some of the problems, like poor mixing and long computation times, one may expect to encounter.

Citation

Download Citation

Tobias Rydén. "EM versus Markov chain Monte Carlo for estimation of hidden Markov models: acomputational perspective." Bayesian Anal. 3 (4) 659 - 688, December 2008. https://doi.org/10.1214/08-BA326

Information

Published: December 2008
First available in Project Euclid: 22 June 2012

zbMATH: 1330.65023
MathSciNet: MR2469793
Digital Object Identifier: 10.1214/08-BA326

Rights: Copyright © 2008 International Society for Bayesian Analysis

JOURNAL ARTICLE
30 PAGES


SHARE
Vol.3 • No. 4 • December 2008
Back to Top