Ensemble Learning Classifiers and Hybrid Feature Selection for Enhancing Intrusion Detection System Performance

Abstract

Network environments must be protected against a variety of cyber-attacks using IDS. The advancement of ML methodologies has yielded notable improvements in intrusion detection system (IDS) capabilities, including greater real-time analysis, adaptability, and accuracy of detection. This study uses machine learning algorithms to give an analytical comparison of several IDS models. The research covers a variety of machine learning approaches, such as supervised and hybrid strategies. We assess these models' performance using important measures including computational efficiency, precision, recall etc. The results show that although supervised machine learning models provides high accuracy, but when used in hybrid model including Random Forests and SVM improves performance. The result is a hybrid model that leverages the strengths of each approach. For instance, Random Forest can provide a robust feature representation, while SVM can refine the decision boundary, leading to a more accurate and reliable classification model. This combination often yields better performance than using any single algorithm alone.