An ensemble of intelligent water drop algorithm for feature selection optimization problem

Abstract

Master River Multiple Creeks Intelligent Water Drops (MRMC-IWD) is an ensemble model of the intelligent water drop, whereby a divide-and-conquer strategy is utilized to improve the search process. In this paper, the potential of the MRMC-IWD using real-world optimization problems related to feature selection and classification tasks is assessed. An experimental study on a number of publicly available benchmark data sets and two real-world problems, namely human motion detection and motor fault detection, are conducted. Comparative studies pertaining to the features reduction and classification accuracies using different evaluation techniques (consistency-based, CFS, and FRFS) and classifiers (i.e., C4.5, VQNN, and SVM) are conducted. The results ascertain the effectiveness of the MRMC-IWD in improving the performance of the original IWD algorithm as well as undertaking real-world optimization problems.