Improving support vector machine using a stochastic local search for classification in datamining

Abstract

In this paper, an improved support vector machine using a stochastic local search (SVM+SLS) is studied for the classification problem in Datamining. The proposed approach tries to find a subset of features that maximizes the classification accuracy rate of SVM. Experiments on some datasets are performed to show and compare the effectiveness of the proposed approach. The computational experiments show that the proposed SVM+SLS provides competitive results and finds high quality solutions. © 2012 Springer-Verlag.