Maximum a Posteriori Estimation for Multivariate Gaussian Mixture Observations of Markov Chains

被引:1334
作者
Gauvain, Jean-Luc [1 ]
Lee, Chin-Hui [2 ]
机构
[1] LIMSI CNRS, Speech Commun Grp, Orsay, France
[2] AT&T Bell Labs, Speech Res Dept, Murray Hill, NJ 07974 USA
来源
IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING | 1994年 / 2卷 / 02期
关键词
Parameter estimation;
D O I
10.1109/89.279278
中图分类号
O42 [声学];
学科分类号
070206 ; 082403 ;
摘要
In this paper, a framework for maximum a posteriori (MAP) estimation of hidden Markov models (HMM) is presented. Three key issues of MAP estimation, namely, the choice of prior distribution family, the specification of the parameters of prior densities, and the evaluation of the MAP estimates, are addressed. Using HMM's with Gaussian mixture state observation densities as an example, it is assumed that the prior densities for the HMM parameters can be adequately represented as a product of Dirichlet and normal-Wishart densities. The classical maximum likelihood estimation algorithms, namely, the forward-backward algorithm and the segmental k-means algorithm, are expanded, and MAP estimation formulas are developed. Prior density estimation issues are discussed for two classes of applications-parameter smoothing and model adaptation-and some experimental results are given illustrating the practical interest of this approach. Because of its adaptive nature, Bayesian learning is shown to serve as a unified approach for a wide range of speech recognition applications.
引用
收藏
页码:291 / 298
页数:8
相关论文
共 31 条
[21]  
LEE CH, 1993, P IEEE INT C AC SPEE, V2, P558
[23]  
PRICE P, 1988, P IEEE INT C AC SPEE, P651
[24]  
PROHOROV YV, 1969, PROBABILITY THEORY
[25]   A SEGMENTAL K-MEANS TRAINING PROCEDURE FOR CONNECTED WORD RECOGNITION [J].
RABINER, LR ;
WILPON, JG ;
JUANG, BH .
AT&T TECHNICAL JOURNAL, 1986, 65 (03) :21-31
[26]  
Rao, 1973, LINEAR STAT INFERENC, V2
[27]   MIXTURE DENSITIES, MAXIMUM-LIKELIHOOD AND THE EM ALGORITHM [J].
REDNER, RA ;
WALKER, HF .
SIAM REVIEW, 1984, 26 (02) :195-237
[28]   EMPIRICAL BAYES APPROACH TO STATISTICAL DECISION-PROBLEMS [J].
ROBBINS, H .
ANNALS OF MATHEMATICAL STATISTICS, 1964, 35 (01) :1-&
[29]  
STERN R, 1987, IEEE T ACOUST SPEECH, V35
[30]  
Zelinski R., 1983, Proceedings of ICASSP 83. IEEE International Conference on Acoustics, Speech and Signal Processing, P1053