A Novel Method of Intrusion Detection Based on Artificial Immune System

This paper presents a novel method of intrusion detection based on artificial immune system. Adopting the constraint-based detectors and any-r intervals matching rule, a novel solution is presented to encode the antibody-antigen. Some immune related concepts are introduced. In order to accelerate the accessing of the normal IP packets, the self-pattern class is proposed. Using the data sets of KDD CUP1999, the experiment results show that the proposed method can achieve a faster running speed and better detecting rates. It also can adapt to dynamically changing environments