A Review of Intrusion Detection Systems: Datasets and machine learning methods

Abstract

At the present time, Security is a crucial issue for all organizations and companies, because intruders are constantly developing new techniques to infiltrate their infrastructure to steal or manipulate sensitive data. Thus, Intrusion Detection System (IDS) has emerged as new technology to protect networks and systems against suspicious activities. Numerous cybersecurity experts highlight the importance of IDS to strength the defensive capacities of systems by alerting for suspicious activities and malicious attacks. Over the years, many techniques like Machine learning (ML) and Deep Learning (DL) have been used to increase the detection accuracy and reduce the false alerts of IDSs. This survey paper presents an overview of some ML and DL algorithms among the most used for IDS. Additionally, because these algorithms depend on the characteristics of malicious events stored in datasets to identify anomalies, we list some publicly available cybersecurity datasets. Furthermore, we highlight the challenges that experts must overcome to enhance the performance of their methods.