Alwan, Hiba Basim and Ku-Mahamud, Ku Ruhana (2013) Feature selection and model selection algorithm using incremental mixed variable ant colony optimization for support vector machine classifier. International Journal of Mathematics and Computers in Simulation, 5 (7). pp. 406-414. ISSN 19980159

PDF Restricted to Registered users only Download (509kB)

Abstract

Support Vector Machine (SVM) is a present day classification approach originated from statistical approaches.Two main problems that influence the performance of SVM are selecting feature subset and SVM model selection. In order to enhance SVM performance, these problems must be solved simultaneously because error produced from the feature subset selection phase will affect the values of the SVM parameters and resulted in low classification accuracy.Most approaches related with solving SVM model selection problem will discretize the continuous value of SVM parameters which will influence its performance.Incremental Mixed Variable Ant Colony Optimization (IACOMV) has the ability to solve SVM model selection problem without discretising the continuous values and simultaneously solve the two problems.This paper presents an algorithm that integrates IACOMV and SVM.Ten datasets from UCI were used to evaluate the performance of the proposed algorithm.Results showed that the proposed algorithm can enhance the classification accuracy with small number of features.

Item Type:	Article
Uncontrolled Keywords:	Incremental Mixed Variable Ant Colony Optimization,Support Vector Machine, Features Selection, Parameter Optimization, Pattern Classification.
Subjects:	Q Science > QA Mathematics > QA76 Computer software
Divisions:	College of Arts and Sciences
Depositing User:	Mrs. Norazmilah Yaakub
Date Deposited:	24 Dec 2013 02:49
Last Modified:	24 Dec 2013 02:49
URI:	https://repo.uum.edu.my/id/eprint/9846

Actions (login required)

View Item