Opposition based gray wolf optimization algorithm for feature selection in classification problems

With the rapid advancement of technology, the data defined in the dataset is distributed and distributed among various classes in order to easily and quickly access the reproduced data. Using the classification algorithms developed to solve classification problems, the data are classified according to similar features. These classification algorithms are trained with the given training dataset, learning is provided, and then they work to classify these data correctly when processing with undetermined test data. Intuitive algorithms have become an increasingly popular algorithm in optimization problems in recent years. Gray Wolf Optimization (GWO) algorithm is a meta heuristic optimization algorithm developed by imitating gray wolves' social and hunting behavior. It was developed by using Gray Wolf optimization algorithm and opposition-based learning method for feature selection of classifiers (K Nearest Neighbour, Support Vector etc.) determined within the scope of this study. Opposite based learning, according to probability theory, the opposite situation of a random point may be closer to the solution than the random point. In the opposition-based learning, the first stage is to determine the opposite-based initial population, and the next stage is the opposition-based generation jump. For the proposed algorithm, innovations such as mutation and boundary value credits have been added apart from opposition-based learning. The opposition-based GWO algorithm developed within the scope of this study was tested simultaneously with the original GWO algorithm for the classification datasets obtained from existing sources and the results were compared. These encounters have been compared for the algorithms' operating times, feature numbers, and accuracy values. The improved GWO algorithm proposed as a result of the comparisons yielded more successful results than the original GWO. Comparisons include time, accuracy, cost value, etc. It was made with factors such as. Keywords Feature selection; Classification; Optimization; Gray Wolf Algorithm; Opposition based learning.