Opposition based Gray Wolf Algorithm for Feature Selection in Classification Problems

Abstract

With the rapid development of technology, the data is classified by distributing among various classes defined on a dataset for easier and faster access to increased data. By using classification algorithms developed for these classification processes, the data is clustered according to similar criteria. These classification algorithms are taught by training with the given training set, and then try to classify the data correctly when the test data that is not given the class is sent. Heuristic algorithms have become very popular in optimization problems in recent years. Gray Wolf Optimization (GWO) algorithm is a meta-heuristic optimization algorithm inspired by the life and hunting strategies of gray wolves in nature. In this study, Gray Wolf Algorithm was developed by using opposition based learning method for different feature selection of certain classifiers (such as KNN & SVM). The proposed improved GWO algorithm was compared with the original GWO algorithm for classification test data sets obtained from the literature.