摘要
本文针对传统的聚类算法在入侵检测系统中的不足,提出一种基于密度的初始聚类中心的选择方法,可克服普通K-Means中的需人工确定K值的问题,用此算法改进的入侵检测模型能够获得很好的聚类效果。对比实验结果,发现使用改进后的算法与传统的K-Means相比可以获得更高的检测率和较低的误报率。
This essay focus on the shortcoming of clustering algorithm on the intrusion detection system,put forward a original clustering center selection based on destiny which can solve the problem of K-Means algorithm need manual set K value. On the basic of this algorithm, The intrusion detection module can make a good effect. After the compared experiments show that the advanced clustering algorithm can improve the efficiency of data clustering.
出处
《科技广场》
2011年第7期95-98,共4页
Science Mosaic
