A novel chaotic chicken swarm optimization algorithm for feature selection

Feature selection is an important task in data mining, which aims to reduce the dimensionality of the data sets while at least maintaining the classification performance. Chicken swarm optimization algorithm (CSO) has been widely applied to feature selection because of its efficiency and effectiveness. However, since feature selection is a challenging task with a complex search space, CSO quickly gets stuck the local minimum problem. This paper aims to improve the CSO searching ability by applying logistic and tend chaotic maps to assist the CSO swarm in exploring the search space better. The proposed chaotic chicken swarm algorithm (CCSO)-based feature selection algorithm is compared with four feature selection algorithms on five benchmark data sets. A comparison among several types of popular classifiers is done to figure out the sensitivity of each classifier corresponding to the selected features and the dimension reduction. During iterations, the best fitness value shows remarkable improvement of the classification accuracy.