Abstract—Feature selection is the process of extracting an optimal subset feature from a primary feature set to minimize data dimensionality. The hybrid metaheuristic is the most common method for dealing with optimization problems. This manuscript proposes a hybrid of Opposition-based Harmony Search (OBHS) and Manta Ray Foraging Optimization (MRFO) for feature selection, which is one of the human-based metaheuristic optimization algorithms. The proposed OBHS-MRFO methodology’s experiments are tested on 21 benchmark datasets taken from the University of California, Irvine (UCI) repository. This dataset is split into three classes: low, medium, and high-scale based, on the dataset dimensions. The proposed model is utilized to overcome the issues of minimum accuracy produced by redundant and irreverent features. The obtained result is compared to four algorithms namely, FS-BGSK, FS-pBGSK, OBHS, and MRFO algorithms. It concludes that the proposed OBHS-MRFO algorithm obtains better results when compared with other methods with regard to average fitness function value, average accuracy, average feature selection size, standard deviation, and computational time.
 
Keywords—feature selection, machine learning, manta ray foraging optimization, metaheuristic algorithm, opposition-based harmony search


Copyright © 2024 by the authors. This is an open access article distributed under the Creative Commons Attribution License (CC BY-NC-ND 4.0), which permits use, distribution and reproduction in any medium, provided that the article is properly cited, the use is non-commercial and no modifications or adaptations are made.