The influence of initial conditions on maximum likelihood estimation of the parameters of a binary hidden Markov model

被引：4

作者：

Dunmur, AP ^{[1
]}

Titterington, DM ^{[1
]}

机构：

[1] Univ Glasgow, Dept Stat, Glasgow G12 8QW, Lanark, Scotland

来源：

STATISTICS & PROBABILITY LETTERS | 1998年 / 40卷 / 01期

关键词：

Baum-Welch algorithm; binary Markov chain; EM algorithm; initial conditions;

D O I：

10.1016/S0167-7152(98)00100-X

中图分类号：

O21 [概率论与数理统计]; C8 [统计学];

学科分类号：

020208 [统计学]; 070103 [概率论与数理统计]; 0714 [统计学];

摘要：

The Baum-Welch (EM) algorithm is a familiar tool for calculation of the maximum likelihood estimate of the parameters in hidden Markov chain models. For the particular case of a binary Markov chain corrupted by binary channel noise a detailed study is carried out of the influence that the initial conditions impose on the results produced by the algorithm. (C) 1998 Elsevier Science B.V. All rights reserved.

引用

页码：67 / 73

页数：7

共 4 条

[1]

ARCHER GEB, 1995, 951 U GLASG DEP STAT

[2]

MAXIMUM LIKELIHOOD FROM INCOMPLETE DATA VIA EM ALGORITHM [J].

DEMPSTER, AP ;

LAIRD, NM ;

RUBIN, DB .

JOURNAL OF THE ROYAL STATISTICAL SOCIETY SERIES B-METHODOLOGICAL, 1977, 39 (01) :1-38

[3]

MAXIMUM-LIKELIHOOD ESTIMATION FOR MULTIVARIATE OBSERVATIONS OF MARKOV SOURCES [J].

LIPORACE, LA .

IEEE TRANSACTIONS ON INFORMATION THEORY, 1982, 28 (05) :729-734

[4]

A TUTORIAL ON HIDDEN MARKOV-MODELS AND SELECTED APPLICATIONS IN SPEECH RECOGNITION [J].

RABINER, LR .

PROCEEDINGS OF THE IEEE, 1989, 77 (02) :257-286

← 1 →