A Review on Network Intrusion Detection System Using Machine Learning

Abstract

After digital revolution, large amount of data are produced from diverse networks from time to time. Hence security of this data is more important. So, there is a need to automate this security system. Intrusion detection systems are considered as the best solution to detect intrusions. Network intrusion detection systems (NIDS) are hired as a defense system to protect networks. Numerous techniques for the development of these defense systems are found in the literature. However, study on the enhancement of datasets used to train and test such security systems is also important. Improved datasets progress the detection capabilities for both offline and online intrusion detection models. Standard datasets like KDD 99, NSL-KDD cup 99 and DARPA 1999 are outdated and they don’t contain data of present attacks such as Denial of Service, therefore they are not suitable for evaluation. In this paper, in depth analysis of CIDDS-001 dataset is shown and the sightings are presented. In this paper, a gist of different papers available related to NIDS are given. This paper even compares all the research papers specifying their merits and demerits. This paper is concluded by providing a research method that can be applied to develop a better network intrusion detection system.