Network Intrusion Prevention by Using Hierarchical Self-Organizing Maps and Probability-Based Labeling

摘要: Nowadays, the growth of computer networks and expansion Internet have made security to be a critical issue. In fact, many proposals for Intrusion Detection/Prevention Systems (IDS/IPS) been proposed. These try avoid that corrupt or anomalous traffic reaches user application operating system. Nevertheless, most IDS/IPS only distinguish between normal can suspected potential attack. this paper, we present approach based on Growing Hierarchical Self-Organizing Maps (GHSOM) which not differentiate but also identify different known attacks. The proposed system has trained tested using well-known DARPA/NSL-KDD datasets results obtained are promising since detect over 99,4% 99,2 % attacker traffic. Moreover, on-line by probability labeling method presented paper.