摘要
本文将多克隆算子和模糊C-均值算法相结合,将多克隆的免疫基因策略作用于特定的亲合度函数来指导聚类过程,得到一种无监督的聚类分析方法。该方法不依赖先验知识、与数据分布无关、能够有效处理大规模原始数据。在KDD CUP99数据集上进行了仿真实验,实验结果表明基于本文方法的入侵检测系统具备良好的性能。
A novel clustering algorithm is developed by using polyclonal algorithm and fuzzy C-means algorithm, which is directed affinity function applied by polyclonal strategy. Independent of transcendental knowledge and data distribution, this algorithm can deal wi] mass unlabeled data, and the computer simulations on the KI）D CUP99 dataset show that this method can achieve good performance. ]
出处
《价值工程》
2013年第25期192-193,共2页
Value Engineering
关键词
多克隆策略
模糊C-均值
入侵检测
oly-clonal strategy
Fuzzy C-Means, intrusion detection
