Comparison on Bayesian Ying-Yang theory based clustering number selection criterion with information theoretical criteria

Author

Lai, Z.B. ; Guo, P. ; Wang, T.J. ; Xu, L.

Author_Institution

Dept. of Comput. Sci. & Eng., Chinese Univ. of Hong Kong, Shatin, Hong Kong

Volume

1

fYear

1998

fDate

4-8 May 1998

Firstpage

725

Abstract

A criterion based on the Bayesian Ying Yang learning theory and system was proposed by Xu (1995, 1996, 1997) for selecting the number of clusters in the clustering analysis and the number of Gaussians in a finite mixture model. In this paper we compare the performance of this criterion with other existing cluster number selection criteria such as Akaike´s information criterion (AIC), CAIC, etc

Keywords

Bayes methods; information theory; learning (artificial intelligence); number theory; statistical analysis; Akaike information criterion; Bayesian Ying-Yang theory; EM algorithm; clustering number selection; information theoretical criteria; learning theory; Bayesian methods; Clustering algorithms; Clustering methods; Computer science; Data analysis; Information analysis; Mean square error methods; Pattern recognition; Statistical analysis; Supervised learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Neural Networks Proceedings, 1998. IEEE World Congress on Computational Intelligence. The 1998 IEEE International Joint Conference on

Conference_Location

Anchorage, AK

ISSN

1098-7576

Print_ISBN

0-7803-4859-1

Type

conf

DOI

10.1109/IJCNN.1998.682370

Filename

682370