An intelligent feature selection and classification method based on hybrid ABC-SVM

Abstract

This paper presents a new approach to feature selection and classifcation based on support vector machine and hybrid artificial bee colony. The approach consists of two stages. At the first stage, this paper presented a hybrid artificial bee colony-based classifier model that combines artificial bee colony to improve classification accuracy with the most superior model parameter and features were selected from the original feature set. The classification accuracy and the feature subset provided by the SVM classifier are both considered to update the food source. Finally, the most superior features and optimal model parameter are fed into SVM to identify different class. The testing results verify the effectiveness of the method in extracting feature subset and pattern classification.