ESTIMATING A MARKOV TRANSITION MATRIX FROM OBSERVATIONAL DATA

Authors
SHERLAWJOHNSON C GALLIVAN S BURRIDGE J
Citation
C. Sherlawjohnson et al., ESTIMATING A MARKOV TRANSITION MATRIX FROM OBSERVATIONAL DATA, The Journal of the Operational Research Society, 46(3), 1995, pp. 405-410
Citations number
7
Categorie Soggetti
Management,"Operatione Research & Management Science","Operatione Research & Management Science
Journal title
The Journal of the Operational Research Society → ACNP
ISSN journal
01605682
Volume
46
Issue
3
Year of publication
1995
Pages
405 - 410
Database
ISI
SICI code
0160-5682(1995)46:3<405:EAMTMF>2.0.ZU;2-3
Abstract
Markov chains are frequently used in Operational Research to describe how a system changes over time, its behaviour being governed by its tr ansition matrix. This paper describes a technique for finding a maximu m likelihood estimate for such a transition matrix when a system is ob served at infrequent time intervals. The technique is called the EM Al gorithm which, for this kind of problem, has distinct advantages over other methods of optimization.