REAL-TIME INTELLIGENT MULTILAYER ATTACK CLASSIFICATION SYSTEM

摘要: Intrusion Detection Systems (IDS) takes the lion’s share of current security infrastructure. intrusions is vital for initiating defensive procedures. detection was done by statistical and distance based methods. A threshold value used in these methods to indicate level normalcy. When network traffic crosses normalcy then above which it flagged as anomalous. there are occurrences new intrusion events increasingly a key part system security, techniques cannot detect them. To overcome this issue, learning helps identifying activities computer system. The objective proposed designed paper classify using an Intelligent Multi Layered Attack Classification System (IMLACS) detecting classifying with improved classification accuracy. intelligent multi layered approach contains three layers. first layer involves Binary Support Vector Machine normal attack. second neural attacks into classes attacks. third fuzzy inference various subclasses. IMLACS can be able behavior networks since better set rules. Feature selection also improve time detection. experimental results show that achieves Rate 97.31%.