Improving SVM Classification Performance with Draco Lizard Feature Selection Optimizer

Abstract

Feature selection plays a vital role in the machine learning process by eliminating redundant and irrelevant attributes from datasets. This not only enhances model accuracy but also reduces training time. When applied to high-dimensional datasets, feature selection helps mitigate the risk of overfitting, thereby enabling the development of models with stronger generalization capabilities. In this study, the recently proposed Draco Lizard Optimizer algorithm is employed for feature selection in conjunction with Support Vector Machines to improve classification performance. The optimization process utilizes an effective fitness function, and experimental evaluations are conducted accordingly. The proposed method is assessed using six different datasets obtained from the UCI Machine Learning Repository. Average number of selected features and classification performance metrics are analyzed in terms of best, worst, average, and standard deviation values. The results demonstrate that the proposed approach yields effective and robust outcomes in feature selection tasks.