摘要
以k-means算法为基础,对入侵检测技术和聚类分析技术进行了研究和分析,指出了传统k-means算法的不足,给出了针对性的改进方法。采用KDDCUP99数据集作为源数据集,对k-means以及改进后的算法进行了仿真实验,实验结果表明,改进后的k-means算法在入侵检测系统中,能够有效地提高入侵检测系统的检测率,降低误报率。
Based on k-means algorithm, the study first analyzed the intrusion detection and clustering techniques, and then pointed out the disadvantages of traditional k-means algorithms as well as the improvements towards them. Adopting KDDCUP99 da- ta sets as source data set, simulation experiments were carried out on both the k-means algorithm and the updated algorithms. The results showed that the updated k-means algorithm can effectively improve the detection rate of the intrusion detection, decrease the probability of false positive ration and achieve the expected performance of the system.
出处
《阜阳师范学院学报（自然科学版）》
2013年第4期80-83,共4页
Journal of Fuyang Normal University(Natural Science)
基金
安徽省教育厅自然科学基金项目(KJ2013Z260)资助
