Intrusion Detection Systems using Supervised Machine Learning Techniques: A survey

• Published in: Procedia computer science Vol. 201; pp. 205 - 212
• Main Authors: Abdallah, Emad E., Eleisah, Wafa’, Otoom, Ahmed Fawzi
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 2022
• Subjects: cyber security; intrusion detection systems; supervised machine learning algorithms; cyber security; supervised machine learning algorithms; intrusion detection systems
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2022.03.029

In this paper, we investigate the subject of intrusion detection using supervised machine learning methods. The main goal is to provide a taxonomy for linked intrusion detection systems and supervised machine learning algorithms. For this purpose, we provide a deep discussion of the concepts of intrusion detection systems, supervised machine learning techniques, and cyber-security attacks. Then, concerning the application of supervised learning for intrusion detection, we cover relevant efforts. Finally, a taxonomy is provided based on these related works. Based on this taxonomy, we can conclude that the classification performance of supervised learning algorithms is high and promising based on a study of four popular data sets in this domain: KDD’99, NSL-KDD, CICIDS2017, and UNSW-NB15. Moreover, feature selection is important and, in many cases, is needed for an enhancement in performance. Furthermore, data imbalance can be a concern, and sampling approaches can help resolve the issue. Finally, for good performance, large intrusion detection data sets necessitate a deep learning technique.