Machine Learning Techniques for Network Intrusion Detection—A Systematic Analysis

Abstract

As the world evolves toward a high dependency on computers and technology, systems and network security became one of the main challenges faced in the last decade. The number of threats that cause potential damage to the network system is rising exponentially due to the increasing complexity of networks and services of modern networks, while situational awareness of the critical assets thus becomes extremely important. The main objective of Cyber security is to protect the electronic data from attacks such as unauthorized network access, intrusion attack, or malware. This study aims to provide an overview of the trends of Machine Learning requirements in supporting the efforts to defend electronic data. The first part of this study is a quantitative analysis of the results obtained from consulting the Scopus and Web of Science databases. In this part, in addition to the most important authors and keywords used, we aimed to find if the countries that are the main targets of cyber-attacks are also involved in researching new ways to improve defending techniques. The second part reveals a brief analysis of Machine learning techniques, risks and innovations and how the next research activity should be conducted, considering the constant evolution of both sides: defenders and attackers.