A Comprehensive Examination of Literature Exploring the Implementation of Machine Learning to Network Security's Intrusion Detection Systems

Abstract

The Internet and telecommunication technologies have developed quickly, the amount of data transferred has greatly increased. Attackers are continually devising new tactics to steal or modify these data because they are so highly desired. The threat these attacks pose to the security of our systems is growing. It is among the most tough issues to resolve for detection of intrusions. An idss is a programme that attempts to analyse network traffic in order to detect intrusions. Despite the fact that many researchers have examined and developed novel IDS systems, IDS even now must be enhanced in order to achieve satisfactory detection capability while reducing number of false alarms. Furthermore, numerous intrusion detection systems have difficulty detecting nil attacks. Machine learning techniques had also recently become popular among scholars as a quick and accurate method of detecting network infiltration. This article offers a taxonomy of machine learning approaches as well as an explanation of IDS. In addition to a list of current IDS that include machine learning and a discussion of the essential components for IDS analysis, this article also outlines the advantages and disadvantages of each machine learning approach. The veracity of the findings from the evaluated study is then discussed after specifics of the various datasets used in the studies are given. The preceding part looks at the results, study obstacles, and projected future trends.