Network Intrusion Detection using Clustering: A Data Mining Approach

Abstract

Network intrusion detection system includes identifying a set of spiteful actions that compromises the basic security requirements such as integrity, confidentiality, and availability of information resources. The enormous increase in network attacks has made the data mining based intrusion detection techniques extremely useful in detecting the attacks. This paper describes a system that is able to detect the network intrusion using clustering concept. This unsupervised clustering technique for intrusion detection is used to group behaviors together depending on their similarity and to detect the different behaviors which are then grouped as outliers. Obviously, these outliers are attacks or intrusion attempts. This proposed method which uses data mining technique will reduce the false alarm rate and improves the security.