Binary spotted hyena optimizer and its application to feature selection

Spotted hyena optimizer (SHO) is a recently developed metaheuristic technique that mimics the hunting behavior of the spotted hyenas. However, it does not provide optimal solution for discrete problems. Therefore, a novel binary version of Spotted Hyena Optimizer is proposed in this paper. The binary version of SHO can deal with discrete optimization problems. In the proposed algorithm, tangent hyperbolic function is utilized to squash the continuous position and then these values are used to update the position of spotted hyenas. The prey searching, encircling, and attacking are three main steps of binary spotted hyena optimizer. The proposed algorithm has been compared with six well-known metaheuristic techniques over 29 benchmark test functions. The effects of convergence, scalability, and control parameters have been investigated. The statistical significance of the proposed approach has also been examined through ANOVA test. The proposed approach is also applied on feature selection domain. The performance of proposed approach has been compared with four well-known metaheuristic techniques over eleven UCI repository datasets. The experimental results reveal that the proposed approach is able to search the optimal feature set than the others.