改进的K-means网络入侵检测算法

Improved K-means Network Intrusion Detection Algorithm

针对K-means算法对于初始聚类中心选择敏感问题,提出了一种改进的K-means算法,该算法优化了聚类中心选择问题,能够获得全局最优的聚类划分,同时减少了算法的时间复杂度。实验结果表明,采用本文的算法进行网络入侵检测,相对于经典的聚类算法,能获得理想的网络入侵检测率和网络误报率。

Aimed at initial cluster center selection sensitivity problem of K-means,an improved K-means is presented in this paper,the algorithm optimizes cluster center selection,which can obtain global optimal cluster partition,at the same time reduce time complexity.The experimental result shows that when the algorithm is used for network detection,compared with the classic cluster algorithm it can get ideal network intrusion detection and false acceptance rate.