An improved support vector machine classifier based on artificial bee colony algorithm

Abstract

Support vector machine (SVM) has unique advantages in the classification of small sample data. The selection of parameters has an important impact on the classification accuracy and generalization ability of SVM. Since the selection of parameters of SVM is usually based on experience, an improved SVM classification model based on artificial bee colony algorithm (GOABC-SVM) is proposed in this paper. In this model, firstly, the traditional artificial bee colony algorithm is optimized using the ideas of global optimal solution guidance and opposite learning. Secondly, we set the reciprocal of classification error rate as the fitness function, and use the improved artificial bee colony algorithm to obtain the optimal parameter combination of SVM. Experiments on a set of datasets of UCI show that the proposed model has higher classification accuracy and better generalization ability.