Artificial immune system based on adaptive clonal selection for feature selection and parameters optimisation of support vector machines. Issue 1 (2nd January 2016)

• Record Type: Journal Article
• Authors: Hashemipour, Maryam Sadat; Soleimani, Seyed Ali
• Is Part Of: Connection science. Volume 28:Issue 1(2016)
• Page Start: 47
• Page End: 62
• Publication Date: 2016-01-02
• Subjects: Adaptive clonal selection algorithm -- artificial immune system algorithm -- feature selection -- support vector machine; Neural computers -- Periodicals; Artificial intelligence -- Periodicals; Cognitive science -- Periodicals; Connectionism -- Periodicals; 006.3
• DOI: 10.1080/09540091.2015.1130022
• Languages: English
• ISSNs: 0954-0091

Abstract

ABSTRACT: Artificial immune system (AIS) algorithm based on clonal selection method can be defined as a soft computing method inspired by theoretical immune system in order to solve science and engineering problems. Support vector machine (SVM) is a popular pattern classification method with many diverse applications. Kernel parameter setting in the SVM training procedure along with the feature selection significantly impacts on the classification accuracy rate. In this study, AIS based on Adaptive Clonal Selection (AISACS) algorithm has been used to optimise the SVM parameters and feature subset selection without degrading the SVM classification accuracy. Several public datasets of University of California Irvine machine learning (UCI) repository are employed to calculate the classification accuracy rate in order to evaluate the AISACS approach then it was compared with grid search algorithm and Genetic Algorithm (GA) approach. The experimental results show that the feature reduction rate and running time of the AISACS approach are better than the GA approach.