FS_KPARD: An effective SVM feature selection method

Author

Wang, Tinghua

Author_Institution

Sch. of Math. & Comput. Sci., Gannan Normal Univ., Ganzhou, China

Volume

2

fYear

2010

fDate

10-12 Aug. 2010

Firstpage

892

Lastpage

895

Abstract

This paper presents an effective feature selection method for support vector machine (SVM). Unlike the traditional combinatorial searching method, feature selection is translated into the model selection of SVM which has been well studied. In more detail, the basic idea of this method is to tune the parameters of the Gaussian ARD (Automatic Relevance Determination) kernel via optimization of kernel polarization, and then to rank all features in decreasing order of importance so that more relevant features can be identified. The proposed method is tested on two UCI data sets to demonstrate its effectiveness.

Keywords

Gaussian processes; feature extraction; learning (artificial intelligence); pattern classification; support vector machines; FS_KPARD; Gaussian ARD; SVM feature selection method; automatic relevance determination; combinatorial searching method; kernel polarization; support vector machine; Accuracy; Correlation; Kernel; Machine learning; Optimization; Support vector machines; Training; auto relevance determination; feature selection; model selection; support vector machine;

fLanguage

English

Publisher

ieee

Conference_Titel

Natural Computation (ICNC), 2010 Sixth International Conference on

Conference_Location

Yantai, Shandong

Print_ISBN

978-1-4244-5958-2

Type

conf

DOI

10.1109/ICNC.2010.5583909

Filename

5583909