An Efficient Swarm based Feature Selection Technique using Random Weight Neural Network

Abstract

Feature selection (FS) is one of the most important pre-processing tasks in machine learning (ML) and data mining, that selects optimum features by eliminating noisy and irrelevant features from the data; to improve the generalization ability of a learning model (i.e., classifier). During the classification process, data with high dimensional feature space requires different optimization techniques to obtain better predictive performance. In this paper we present a swarm intelligence based technique called binary artificial bee colony (Binary-ABC) to obtain optimum feature subset. Different binary and multiclass datasets are utilized to evaluate the performance of our proposed technique. Experimental results show that our technique provides better generalization ability with random weight neural network (RWNN), when compare with other ML classifiers.