基于PCA与R-连续位匹配的亲和力改进算法

IMPROVED AFFINITY ALGORITHM BASED ON PCA AND R-CONTIGUOUS BIT MATCHING

人工免疫算法中传统亲和力计算方法一般采用r-连续位匹配规则,但这种匹配规则不能有效识别已知入侵的变种,且编码长度直接影响入侵检测的效率。为提高检测率,在PCA特征提取的基础上,采用分段加权的思想改进亲和力算法。实验结果表明,改进的亲和力算法能有效提高入侵检测的检测率。

Traditional affinity calculation method in artificial immune algorithm usually uses r-contiguous bit matching rules,but it can not effectively identify the variants of the known intrusions,and the encoding length directly impacts the efficiency of intrusion detection.In order to enhance the detection efficiency,in this paper we use PCA to extract invasion features and then improve affinity algorithm with the idea of sub-weighted.Experimental results show that the improved algorithm can effectively raise intrusion detection efficiency.