Application of artificial immune system algorithms for intrusion detection

Abstract

Intrusion Detection (ID) is one of the most challenging problems in today’s era of computer security. New innovative ideas are used by the hackers to break the security, hence the challenge for developing better ID systems are increasing day-by-day. In this paper, we applied the Artificial Immune System (AIS) based classifiers for intrusion detection. Each classifier is evaluated based on high accuracy and detection rate with low false alarm rate. The results are compared using percentage split (80%) and cross validation (10 fold) test options basing on two nominal target attributes i.e. type of attacks and protocol types having 5 and 3 sub-classes respectively. The experimental results indicate that the performance of CSCA (clonal selection classification algorithm) is better AIS based classifier for network based Intrusion Detection.