Quantum based Whale Optimization Algorithm for wrapper feature selection

• Published in: Applied soft computing Vol. 89; p. 106092
• Main Authors: Agrawal, R.K., Kaur, Baljeet, Sharma, Surbhi
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.04.2020
• Subjects: Bio-inspired technique; Evolutionary techniques; Feature selection; Quantum; Swarm based techniques; Whale Optimization Algorithm; Quantum; Whale Optimization Algorithm; Feature selection; Evolutionary techniques; Bio-inspired technique; Swarm based techniques
• ISSN: 1568-4946; 1872-9681
• DOI: 10.1016/j.asoc.2020.106092

In this paper, we propose the Quantum Whale Optimization Algorithm (QWOA) for feature selection, which is an amalgamation of the Quantum Concepts and the Whale Optimization Algorithm (WOA). The proposed method enhances the exploratory and exploitation power of the classical WOA, with the use of quantum bit representation of the individuals of the population and the quantum rotation gate operator as a variation operator. Modified mutation and crossover operators are also introduced for quantum-based exploration, shrinking and spiral movement of the whales in the proposed QWOA. The efficacy of the proposed method is compared with that of the conventional WOA and with well-known evolutionary, swarm and quantum algorithms with fourteen datasets from diversified domains. Experimental results demonstrate the superior performance of the proposed QWOA method. Statistical tests also demonstrate the significantly better performance of the QWOA in comparison to eight well-known meta-heuristic algorithms. •A quantum-based approach of the WOA is proposed using the quantum-bit representation.•Modified mutation, crossover and quantum rotation gate operators are introduced.•The performance compared with WOA, GA, PSO, BBA, GWO, SSA, QGA and QPSO.•Results on 14 datasets in terms of fitness, accuracy, AUC and the number of features.•Statistical tests performed.