Experimental Analysis of Data Mining Application for Intrusion Detection with Feature Reduction

Abstract: The reliability and availability of network services is under threat as Denial-of-Service (DoS) attacks develop. It needs efficient mechanisms for detecting DoS attacks. Investigate and derive second- order information from traffic data found on the network. Such second-order statistics derived from the proposed approach to analysis may provide valuable correlative information that is concealed among the apps. Through using this secret information, the accuracy of detection can be significantly improved. Comparisons also show that our Cyber Crime based detection approach by Applying Data Mining techniques outperforms some other existing DoS attack detection work.

Keywords: Denial- of-Service (DoS), TCP & UDP, Local to User (R2L), Network Intrusion Detection Systems (NIDS).