期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2015
卷号：8
期号：11
页码：101-112
DOI：10.14257/ijhit.2015.8.11.08
出版社：SERSC
摘要：In this paper, a novel negative selection algorithm for recognition problems was given. Compared with the traditional negative selection algorithm, a co-stimulation signal was added to start the detectors, which a key factor in immune response. Co-stimulation signal was calculated by the techniques of the statistics and the sliding window, which not only reduced time complexity of algorithm but also improved accuracy of the algorithm. Entropy was adopted to evaluate the density of detectors for optimizing the coverage of nonself area. Experiment results proved high accuracy and efficiency of the proposed algorithm.
