Research on the Performance of Machine Learning Algorithms for Intrusion Detection System

Abstract

Intrusion Detection System (IDS) is a critical approach to ensure network system security. Currently, network attacks are complicated and volatile. Moreover, hackers are also more inclined to adopt new attack techniques to obtain users’ privacy. Under this circumstance, the intelligent intrusion detection system has become the primary approach to detect hackers’ attacks. In this study, intelligent intrusion detection models applying the novel UNSW-NB15 data set as well as various machine learning algorithms are investigated. Furthermore, data set is pre-processed through using one-hot encoding and normalization in our experiment. Subsequently, the performance comparison of six different types of machine learning algorithms in intrusion detection tasks was implemented. The experimental results reveal that in the complex and changeable network traffic data, machine learning technology has presented desired performance.