Extensive Analysis of Intrusion Detection System Using Deep Learning Techniques

Intrusion detection systems (IDS) is a major cyber security approach that aims to observe the status of the software and hardware components operating in a system or network. Several IDS models have been available in the literature for tackling the security issues, which can be divided into two types, namely, signature-based IDS (SIDS) and anomaly-based IDS (AIDS). Regardless of recent developments, the present IDSs still needed to enhance the detection performance, minimize the false alarms, and recognizing unknown attack. For resolving these issues, several research works have dedicated on the design of IDS via machine learning (ML) and deep learning (DL) models. In this aspect, this study intends to perform a complete review of recently developed DL models for IDS. Besides, a detailed review of various DL models designed to identify the intrusions in the network take place. In addition, an extensive analysis of the reviewed approaches is performed in terms of different aspects such as objectives, underlying methodology, dataset used, and measures. Moreover, a brief discussion of the results obtained by the DL-based IDS models is also made. At last, the possible future developments and challenges involved in the IDS models are elaborated briefly.