Determination of the number of components based on class separability in mixture-based classifiers

Author

Tenmoto, H. ; Kudo, Mineichi ; Shimbo, Masaru

Author_Institution

Dept. of Inf. Eng., Kushiro Nat. Coll. of Technol., Japan

fYear

1999

fDate

36495

Firstpage

439

Lastpage

442

Abstract

We propose a novel method for determining the number of components in mixture-based classifiers. Each class-conditional probabilistic density function can be approximated well by the mixture of Gaussian components. However, the performance of this classifier depends on the number of components. In our proposed method, determination of the number of components is based on both probabilistic likelihood and class separability. The results of experiments confirmed the effectiveness and the property

Keywords

Gaussian processes; data handling; pattern classification; probability; Gaussian components; class separability; class-conditional probabilistic density function; minimum description length; mixture-based classifiers; pattern recognition; probabilistic likelihood; Australia; Bayesian methods; Clustering algorithms; Covariance matrix; Density functional theory; Educational institutions; Gaussian distribution; Intelligent systems; Pattern recognition; Systems engineering and theory;

fLanguage

English

Publisher

ieee

Conference_Titel

Knowledge-Based Intelligent Information Engineering Systems, 1999. Third International Conference

Conference_Location

Adelaide, SA

Print_ISBN

0-7803-5578-4

Type

conf

DOI

10.1109/KES.1999.820217

Filename

820217