Advancing Network Intrusion Detection Systems with Machine Learning Techniques

Abstract

This paper presents an approach to enhancing the efficiency and effectiveness of Network Intrusion Detection Systems (NIDS) by leveraging Machine Learning (ML) techniques, specifically Decision Trees (DT), Naïve Bayes (NB), and Support Vector Machine (SVM). The pro-posed methodology involves a comprehensive evaluation and comparison of these algorithms using the NSL-KDD and UNSW-NB15 datasets, employing standard evaluation metrics such as accuracy, precision, recall, and F1-score. The study identifies the most effective algorithm for practical NIDS deployment. By providing actionable insights and recommendations for implementing the most suitable ML algorithm, this research contributes significantly to the ongoing efforts in strengthening network security against evolving cyber threats.