An Intelligent Network Intrusion Detection Using Data Mining Techniques

Abstract

Network Intrusion Detection is to detect malicious attacks to the networks for different uses from military to enterprise. Currently available approaches either rely on the known network attacks or have high proportion of normal network traffics that were erroneously reported as anomalous traffics. The aim of this paper is to develop an efficient algorithm for intrusion detection without prior knowledge of network attacks. Uniquely, our approach will integrate a newly developed data mining technique for data feature classification with techniques commonly used for human detection. The key idea is to achieve on-line and automated learning of new attacks for precise and real-time intrusion detection.