Bayesian online algorithms for learning in discrete Hidden Markov Models

Abstract

We propose and analyze two different Bayesian online algorithms for learning in discrete Hidden Markov Models and compare their performance with the already known Baldi-Chauvin Algorithm. Using the Kullback-Leibler divergence as a measure of generalization we draw learning curves in simplified situations for these algorithms and compare their performances.

Divisions: College of Engineering & Physical Sciences > Systems analytics research institute (SARI)
Additional Information: This is a pre-copy-editing, author-produced PDF of an article accepted for publication in Discrete and Continuous Dynamical Systems Series B following peer review. The definitive publisher-authenticated version Alamino, Roberto C. and Caticha, Nestor (2008) Bayesian online algorithms for learning in Hidden Markov Models. Discrete and Continuous Dynamical Systems Series B , 9 (1). pp. 1-10. ISSN 1531-3492 is available online at: http://aimsciences.org/journals/pdfs.jsp?paperID=2980&mode=abstract
Uncontrolled Keywords: Bayesian online algorithms,discrete Hidden Markov Models,Baldi-Chauvin algorithm,Kullback-Leibler divergence,learning curves,Applied Mathematics,Discrete Mathematics and Combinatorics
Publication ISSN: 1553-524X
Last Modified: 04 Nov 2024 08:09
Date Deposited: 09 Feb 2010 13:56
Full Text Link:
Related URLs: http://www.scop ... tnerID=8YFLogxK (Scopus URL)
http://aimscien ... 0&mode=abstract (Publisher URL)
PURE Output Type: Article
Published Date: 2008-01
Authors: Alamino, Roberto C. (ORCID Profile 0000-0001-8224-2801)
Caticha, Nestor

Download

[img]

Version: Accepted Version


Export / Share Citation


Statistics

Additional statistics for this record