Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
SMEM algorithm for mixture models
Autore:
Ueda, N; Nakano, R; Ghahramani, Z; Hinton, GE;
Indirizzi:
NTT Commun Sci Labs, Seika, Kyoto 6190237, Japan NTT Commun Sci Labs Seika Kyoto Japan 6190237 Seika, Kyoto 6190237, Japan Univ Coll London, Gatsby Computat Neurosci Unit, London WC1N 3AR, England Univ Coll London London England WC1N 3AR Unit, London WC1N 3AR, England
Titolo Testata:
NEURAL COMPUTATION
fascicolo: 9, volume: 12, anno: 2000,
pagine: 2109 - 2128
SICI:
0899-7667(200009)12:9<2109:SAFMM>2.0.ZU;2-J
Fonte:
ISI
Lingua:
ENG
Soggetto:
EM ALGORITHM;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Life Sciences
Engineering, Computing & Technology
Citazioni:
16
Recensione:
Indirizzi per estratti:
Indirizzo: Ueda, N NTT Commun Sci Labs, Seika, Kyoto 6190237, Japan NTT Commun Sci Labs Seika Kyoto Japan 6190237 yoto 6190237, Japan
Citazione:
N. Ueda et al., "SMEM algorithm for mixture models", NEURAL COMP, 12(9), 2000, pp. 2109-2128

Abstract

We present a split-and-merge expectation-maximization (SMEM) algorithm to overcome the local maxima problem in parameter estimation of finite mixturemodels. In the case of mixture models, local maxima often involve having too many components of a mixture model in one part of the space and too few in another, widely separated part of the space. To escape from such configurations, we repeatedly perform simultaneous split-and-merge operations using a new criterion for efficiently selecting the split-and-merge candidates. We apply the proposed algorithm to the training of gaussian mixtures and mixtures of factor analyzers using synthetic and real data and show the effectiveness of using the split-and-merge operations to improve the likelihoodof both the training data and of held-out test data. We also show the practical usefulness of the proposed algorithm by applying it to image compression and pattern recognition problems.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 22/01/20 alle ore 06:56:57