Skip to Main Content
The negative selection algorithm presented in this paper is inspired by the mechanism exhibited in biological immune T cells negative selection process in thymus. The algorithm is made up of three procedures: definition of self space, generation of detectors, monitor the variance of self set. The real valued representation, which is used in this paper, is closer to the original problem space. It adopts the concept of detection radius and reduces data redundancy in the detector set. It has few parameters and stable. This paper analyses the application of the algorithm on anomaly detection and the results show that the algorithm has fascinating ability on anomaly detection.