Employing the Social Group Optimization Algorithm to Find the Best Hyper parameters of Support Vector Machine

Abstract

Support Vector Machine (SVM) is one of the most widely used algorithms for solving classification and regression problems. SVM parameters such as the kernel and the penalty (C) parameters greatly affect the classification accuracy. For the purpose of improving classification accuracy within a record implementation time, a Social Group Optimization (SGO) algorithm was proposed to find the best combination of SVM parameters through which to improve the performance and to obtain the highest classification accuracy and speed of execution. Five different types of datasets (Iris, Wine, Glass, Stat log and Car) were used that all of them were taken from the (UCI) repository. Moreover, a medical dataset, which is taken from (Dread, Alia, 2019), was used to verify the proposed algorithm. The results of the proposed algorithm were compared with the Grid Search algorithm (GS). The comparison results showed a preference for the (SGO) algorithm compared to the (GS) algorithm in terms of classification accuracy and speed of implementation for all the used datasets in this work.