Study on Adaptive ID Modeling Based on Immune Principles

This paper proposes a novel Immune-based adaptive intrusion detection model (IAIDM). In the model, a minimally complete detector repertoire is firstly specified to avoid the heavy iterative process for detector generation in traditional immune mode. Meanwhile, it provides better characterization of the boundary between self space and nonself space with density of cells. Secondly, a mechanism of abnormity presenting and abnormity triggering is provided to generate detectors only when need, which helps to flexible and adaptive detection. Lastly, the evolution of detectors is designed to realize dynamic update and associative memory. Experiment results show the modeling approach proposed for ID systems is feasible.