An entropy criterion for assessing the number of clusters in a mixture model

被引：1759

作者：

Celeux, G

Soromenho, G

机构：

[1] UNIV LISBON, LEAD, P-1600 LISBON, PORTUGAL

[2] INST NATL RECH INFORMAT & AUTOMAT RHONE ALPES, F-38330 Montbonnot St Martin, FRANCE

来源：

JOURNAL OF CLASSIFICATION | 1996年 / 13卷 / 02期

关键词：

cluster analysis; Gaussian mixture; entropy; Bayesian criteria;

D O I：

10.1007/BF01246098

中图分类号：

O1 [数学];

学科分类号：

0701 ; 070101 ;

摘要：

In this paper, we consider an entropy criterion to estimate the number of clusters arising from a mixture model. This criterion is derived from a relation linking the likelihood and the classification likelihood of a mixture. Its performance is investigated through Monte Carlo experiments, and it shows favorable results compared to other classical criteria.

引用

收藏

页码：195 / 212

页数：18

相关论文

共 33 条

[1] MIXTURE-MODELS, OUTLIERS, AND THE EM ALGORITHM [J].

AITKIN, M ;

WILSON, GT .

TECHNOMETRICS, 1980, 22 (03) :325-331

[2]

AITKIN M, 1985, J ROY STAT SOC B MET, V47, P67

[3] NEW LOOK AT STATISTICAL-MODEL IDENTIFICATION [J].

AKAIKE, H .

IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1974, AC19 (06) :716-723

[4] MODEL-BASED GAUSSIAN AND NON-GAUSSIAN CLUSTERING [J].

BANFIELD, JD ;

RAFTERY, AE .

BIOMETRICS, 1993, 49 (03) :803-821

[5]

Bezdek J. C., 1981, Pattern recognition with fuzzy objective function algorithms

[6]

Bock H.H., 1989, CONCEPTUAL NUMERICAL, P12

[7] ON SOME SIGNIFICANCE TESTS IN CLUSTER-ANALYSIS [J].

BOCK, HH .

JOURNAL OF CLASSIFICATION, 1985, 2 (01) :77-108

[8] MULTI-SAMPLE CLUSTER-ANALYSIS USING AKAIKES INFORMATION CRITERION [J].

BOZDOGAN, H ;

SCLOVE, SL .

ANNALS OF THE INSTITUTE OF STATISTICAL MATHEMATICS, 1984, 36 (01) :163-180

[9] ON THE INFORMATION-BASED MEASURE OF COVARIANCE COMPLEXITY AND ITS APPLICATION TO THE EVALUATION OF MULTIVARIATE LINEAR-MODELS [J].

BOZDOGAN, H .

COMMUNICATIONS IN STATISTICS-THEORY AND METHODS, 1990, 19 (01) :221-278

[10]

Bozdogan H., 1993, Information and classification: Concepts, methods and applications, P40, DOI DOI 10.1007/978-3-642-50974-2_5

← 1 2 3 4 →