Simultaneous Feature Selection and LS-SVM Parameters Optimization Algorithm Based on PSO

Author

Yao, Quan-Zhu ; Cai, Jie ; Zhang, Jiu-long

Author_Institution

Sch. of Comput. Sci. & Eng., Xi´´an Univ. of Technol., Xi´´an, China

Volume

5

fYear

2009

fDate

March 31 2009-April 2 2009

Firstpage

723

Lastpage

727

Abstract

For the feature selection and parameter optimization of LS-SVM, propose a At first, a population of particles (feature subsets) was randomly generated, then the features and parameters are optimized by PSO algorithm. The experiments on the UCI database indicate that the proposed method can efficiently find the suitable feature subsets and LS-SVM parameters. Also, comparison are made against GALS-SVM and LS-SVM; and the results show that the proposed PSOLS-SVM outperform the others in classification performance.

Keywords

least squares approximations; particle swarm optimisation; pattern classification; support vector machines; feature selection; least squares support vector machine; parameter optimization; particle swarm optimization algorithm; pattern classification; Computer aided instruction; Computer science; Equations; Fuses; Least squares methods; Machine learning; Particle swarm optimization; Spatial databases; Support vector machine classification; Support vector machines; LS-SVM; feature selection; parameters optimization; partical swarm optimization algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Science and Information Engineering, 2009 WRI World Congress on

Conference_Location

Los Angeles, CA

Print_ISBN

978-0-7695-3507-4

Type

conf

DOI

10.1109/CSIE.2009.148

Filename

5170628