Optimal Feature Selection Using Modified COVID Optimization Algorithm

Various data dimensions are the most significant challenge today due to the rapid development and increase of data. It is necessary to perform data mining techniques to get the desired information from this pile of data. One part of data mining is data pre-processing, namely feature selection. Feature selection is the process of selecting irrelevant features that are considered to worsen the performance of the data classification process and computation time. In this paper, we set the optimal feature subset with a wrapper technique using a KNN classifier to detect the best feature subset simultaneously using a modified COVID optimization algorithm using the concept of competition strategy as a selection operator to select the relevant feature subset from different data dimensions. The authors’ experimental results show that the performance of the proposed algorithm is superior to the binary version of the algorithm in handling the selection of relevant features, thus achieving a better accuracy rate than before.