IMPROVING USER-TO-ROOT AND REMOTE-TO-LOCAL ATTACKS USING GROWING HIERARCHICAL SELF ORGANIZING MAP

Intrusion Detection System (IDS) protects a system by detecting “known” as well as “unknown” attacks and generates the alert for suspicious activities in the traffic. There are various approaches for IDS, but our survey was focused on IDS using Self Organizing Map (SOM). Our survey shows that the existing IDS based on Self Organizing Map (SOM) have more computational time and poor detection rate for User-to-Root (U2R) and Remoteto-Local (R2L) attacks. So, our objective is to improve the detection rate of U2R and R2L attacks along with low computational time. From our survey we found that, Growing Hierarchical Self Organizing Map (GHSOM) is efficient due to its low computation time compared to traditional SOM model. To achieve our objective, our model uses GHSOM algorithm along with proper features selection to improve the performance of U2R and R2L attacks. Our empirical result indicates that, there is nearby 75% increase in the detection rate of U2R and R2L attacks by using GHSOM approach compare to SOM approach.