Negative selection algorithm: A survey on the epistemology of generating detectors

Abstract

Within the Artificial Immune System community, the most widely implemented algorithm is the Negative Selection Algorithm. Its performance rest solely on the interaction between the detector generation algorithm and matching technique adopted for use. Relying on the type of data representation, either for strings or real-valued, the proper detection algorithm must be assigned. Thus, the detectors are allowed to efficaciously cover the non-self space with small number of detectors. In this paper, the different categories of detection generation algorithm and matching rule have been presented. Briefly, the biologial and artificial immune system, as well as the theory of negative selection algorithm were introduced. The exhaustive detector generation algorithm used in the original Negative Selection Algorithm laid the foundation at profer-ring other algorithmic methods based on set of rules in generating valid detectors for revealing anomalies. © Springer Science+Business Media Singapore 2014.