A new hybrid algorithm based on binary gray wolf optimization and firefly algorithm for features selection
Omar Saber Qasim, Noor Muhammed Noori · Asian-European Journal of Mathematics · 2020
In classifying diseases, many datasets have a number of redundant features that do not affect classification accuracy. There are several evolutionary algorithms that are used to determine the feature and reduce dimensional patterns such as the gray wolf optimization (GWO) and the firefly algorithm (FFA). In this paper, a hybrid optimization algorithm BGWO_FFA was proposed to find the optimal subset of datasets. BGWO_FFA benefits from the ability of both GWO and FFA to find the best subset of features in search space. The results show that BGWO_FFA significantly outperformed the results of the algorithm BGWO as it showed efficient and high accuracy through the mean squared error (MSE) and the number of features selection.