A Hybrid Binary Grey Wolf Optimizer Based on DE Algorithm for Feature Selection

Feature selection is one kind of optimization problem that has bio-objective functions, where it is necessary to get the minimum number of features that achieve high classification accuracy. According to literature studies, several kinds of meta-heuristic algorithms have been utilized to solve feature selection problems. One of these meta-heuristic algorithms is the Grey Wolf Optimization (GWO) algorithm and its modified variants, including the binary versions. They have yielded competitive results compared to other algorithms. Despite the simplicity and effectiveness of GWO and its modified versions, they face challenges related to the exploitation ability of the local search. To avoid premature convergence as well as limited refinement around promising solutions, further improvements must be included. In our paper, an enhanced hybrid algorithm, named BGWODE, is proposed to improve the performance of the binary GWO (BGWO) in wrapper-based techniques for solving feature selection problems. The proposed BGWODE algorithm is developed by hybridizing the BGWO and differential evolution (DE) algorithms. Incorporating DE improves the search capability of BGWO in terms of exploitation. To evaluate the effectiveness of the proposed BGWODE algorithm, a total of 18 well-known feature selection datasets from UCI Repository were employed. A comparison was conducted between BGWODE and six famous state-of-the-art algorithms. The classification process is done using the k-nearest-neighbour (KNN) classifier. The results demonstrated the ability of BGWODE to achieve high classification accuracy with a reduced number of selected features across the majority of benchmark datasets used in the comparison. For the majority of the tested datasets, BGWODE was able to attain high level of classification accuracy with few numbers of selected features.