Genetic Operator based Binary Chaotic Grey Wolf Optimizer for Feature Selection

Abstract

In today's world, enormous data is generated through various IT devices. It is tedious and complex to process these data for extracting useful information. Feature selection (FS) plays a meaningful role in selection of highly relevant features from high dimensional dataset. The presented paper introduces a wrapper-based FS technique that integrates chaotic map (CM), genetic crossover operator (CR) and mutation(M) with binary Grey Wolf optimizer (BGWO) called CBGWO, to tackle FS problem. To increase the efficiency of the proposed approach for the FS domain, three crossover operators have been used: one-point, two-point, and uniform. To validate the effectiveness of proposed method eleven well-known UCI datasets is utilized for experiment and comparison is carried out with other well known literature.The results obtained from the proposed methods has optimal performance than the other baseline appraoches, especially when using binary chaotic GWO with uniform crossover.