A Preliminary Study on the Application of Hybrid Machine Learning Techniques in Network Intrusion Detection Systems

Abstract

This study explores the application of hybrid machine learning techniques in the domain of network intrusion detection systems (NIDS). The traditional approach to network intrusion detection typically involves the use of rule-based systems or signature-based systems. Rule-based systems use a set of predefined rules to detect known attack patterns, while signature-based systems use a database of known attack signatures to match against incoming network traffic. While these approaches can be effective at detecting known attacks, they are often not effective at detecting novel or unknown attacks. This is because rule-based and signature-based systems rely on pre-defined rules or signatures and may not be able to identify new or previously unseen attack patterns. Hybrid machine learning techniques were developed in response to these limitations. The authors review and compare recent studies that combine multiple machine learning algorithms and techniques to enhance the accuracy and efficiency of NIDS. The authors conclude that hybrid machine learning techniques are effective in improving the accuracy and reducing the false positives of NIDS. The study highlights the potential of hybrid techniques in enhancing the performance of NIDS, which is crucial in detecting and preventing cyber-attacks in various organizations and critical infrastructures.