Modified-MCA: An Effective Feature Selection

Abstract

Feature selection is often an essential data processing step prior to applying a learning algorithm. The removal of irrelevant and redundant information improves the performance of machine learning algorithms. In this paper, Modified-Multiple Correspondence Analysis (M-MCA) is proposed. It explores the correlation between different features and classes to score the features for feature selection. To evaluate the performance of proposed Modified-MCA, experiments are carried out on ten benchmark datasets. In the experiments, AdaBoost, Decision Table, JRip, Naïve Bayes, and Sequential Minimal Optimization (SMO) are used as the classifiers. The proposed Modified-MCA demonstrates the promising classification results and performs better than other well-known feature selection methods; Information Gain and Relief.