Knowledge Based Analysis of Statistical Tools in Attack Detection
With the network playing more and more important effect in the modern society, crimes using computer network are presenting the obvious trend of escalation. In order to ensure network security, one technique adopted for detecting abnormal or unauthorized behavior is the Intrusion Detection System (IDS). Various data mining techniques are applied in an offline environment to add more depth to the network defense in order to determine the various attacks or threats to the network. This paper focuses on finding the best classifier with respect to accuracy in detecting various attacks on the tcpdump data so that mechanisms can be incorporated in Intrusion Detection Systems to detect the misclassified types of attacks efficiently.