Comments on "Efficient training algorithms for HMM's using incremental estimation"

被引：6

作者：

Byrne, W ^{[2
]}

Gunawardana, A

机构：

[1] Johns Hopkins Univ, Dept Elect & Comp Engn, Baltimore, MD 21218 USA

[2] Johns Hopkins Univ, Ctr Language & Speech Proc, Baltimore, MD 21218 USA

来源：

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING | 2000年 / 8卷 / 06期

关键词：

EM algorithm; hidden Markov models; incremental estimation; maximum likelihood estimation;

D O I：

10.1109/89.876315

中图分类号：

O42 [声学];

学科分类号：

070206 ; 082403 ;

摘要：

Efficient training algorithms for HMM's using incremental estimation [1] investigates EM procedures that increase training speed. The authors' claim that these are GEM [2] procedures Ts incorrect. We discuss why this is so, provide an example of nonmonotonic convergence to a local maximum in Likelihood, and outline conditions that guarantee such convergence.

引用

页码：751 / 754

页数：4

共 13 条

[1]

[Anonymous], 1998, LEARNING GRAPHICAL M

[2] Accelerating the EMML algorithm and related iterative algorithms by rescaled block-iterative methods [J].

Byrne, CL .

IEEE TRANSACTIONS ON IMAGE PROCESSING, 1998, 7 (01) :100-109

[3]

BYRNE W, 1999, EUR C SPEECH COMM TE

[4] 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

[5] SPACE-ALTERNATING GENERALIZED EXPECTATION-MAXIMIZATION ALGORITHM [J].

FESSLER, JA ;

HERO, AO .

IEEE TRANSACTIONS ON SIGNAL PROCESSING, 1994, 42 (10) :2664-2677

[6] Efficient training algorithms for HMM's using incremental estimation [J].

Gotoh, Y ;

Hochberg, MM ;

Silverman, HF .

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 1998, 6 (06) :539-548

[7]

GUNAWARDANA A, CONVERGENCE EM VARIA

[8]

Hart P.E., 1973, Pattern recognition and scene analysis

[9]

HERO A, 1998, IEEE SIGNAL PROCESSI, P24

[10] ACCELERATED IMAGE-RECONSTRUCTION USING ORDERED SUBSETS OF PROJECTION DATA [J].

HUDSON, HM ;

LARKIN, RS .

IEEE TRANSACTIONS ON MEDICAL IMAGING, 1994, 13 (04) :601-609

← 1 2 →