Intrusion Detection System Using Bayesian Network and Feature Subset Selection

Abstract

Attack detection is a challenging area of research in the field of information security. Intruders use various techniques to gain the unauthorized access to computer network. Intrusion detection is a mechanism for detecting and preventing various attacks. Recently Intrusion Detection System (IDS) along with antivirus software plays a vital role in information security architecture of many organizations. Many machine learning approaches have been used to increase the efficiency and detection rate of intrusion detection system. Irrelevant and redundant features in data will cause a problem in network traffic classification. These irrelevant and redundant features will slow down the process of classification and also prevent a classifier from making accurate and efficient decisions when handling with big data. In this paper, we propose intrusion detection system using Bayesian network and feature selection. The performance is evaluated using Kyoto data set. The Experimental results show that our proposed approach achieves with a detection rate of 99.9% and efficiency in detecting network traffic attack.