摘要
提出基于克隆选择的模糊聚类算法,将该聚类算法用于网络入侵检测,针对入侵数据的混合属性改进距离测度的计算方法,实现了对大规模混合属性原始数据的异常检测,并能有效检测到未知攻击。在KDD CUP99数据集中进行了对比仿真实验,实验结果表明算法对已知攻击和未知攻击的检测率以及算法的误警率都是理想的。
The clustering algorithm is employed for the network to detect the intrusions in this paper. And in order to treat the data set with mixed numeric and categorical values, a novel algorithm for mixed data by modifying the common cost function and race of the within cluster dispersion matrix is used here. So the intrusion detection system can deal with mass unlabeled data to distinguish between normal and anomaly and to detect unknown attacks effectively. The simulations on the KDD CUP99 dataset show that the detection rate of known attacks and unknown attacks and the false positive rate of this algorithm are excellent.
出处
《微电子学与计算机》
CSCD
北大核心
2007年第3期135-137,141,共4页
Microelectronics & Computer
基金
航空科学基金项目(05F53032)
关键词
克隆选择算法
聚类分析
入侵检测
clonal selection
cluster analysis
intrusion detection
