Performance Evaluation of Artificial Immune System based Classifiers in Intrusion Detection

摘要

An important component of Intrusion Detection Systems (IDS) is pattern recognition and classification. Intrusion detection systems must deal with intrinsic characteristics that are difficult to detect using classical methods. One of the major challenges of IDS is the high asymmetry between the normal and the abnormal state. Human Immune System (HIS) successfully protects the body against a vast variety of foreign pathogens and faces some of the common problems faced in IDS. It has to recognize patterns and classify on highly asymmetric information. For an HIS to be successful, it has to be self-adaptive and self-learning. In this paper we investigate Artificial Immune System (AIS) which mimic the HIS and its effectiveness in IDS.