A Novel Feature Selection Method for Support Vector Machines Using a Lion's Algorithm

摘要

Feature selection is an effective method on the solving of classification problem. Through reduce unnecessary features, feature selection can improve accuracy of classification and reduce the training time of classifiers. Besides, the bio-inspired algorithms are usually used to search the optimal feature subset for a classifier. This study adopted the newly developed bio-inspired algorithm with social behavior of lions to select optimal feature subset for support vector machines. By using the dataset of UCI machine learning database, the proposed method was compared with the genetic algorithms. Experimental results demonstrated that the performance of the proposed method was superior to that of genetic algorithms.