摘要
为了进一步提高网络入侵检测技术的检测率,降低误报率和漏报率。针对普通聚类算法存在的聚类结果对随机选取初始聚类中心敏感、分类结果不稳定,从而造成的检测率低、漏报和误报率高的特点。提出一种基于动态聚类算法的网络入侵检测模型,实验结果表明通过在K-均值聚类算法的基础上增加动态迭代调整聚类中心,使聚类结果更稳定更准确。与K-均值聚类等算法相比提高了网络入侵检测的性能,从而表明该算法的可行性、有效性。
In order to improve the detection rate,lower false alarm rate and missing rate, further. A network intrusion detection model that combines a dynamic clustering algorithm and intrusion detection technology is proposed. The general clustering algorithm is sensitive of the initial cluster center using, and it can make the classification results of instability, resUlting in the detection rate is low,and failing to report the characteristics of a high false alarm rate. A model of network intrusion detection based on dynamic clustering algorithm is introduced. Experimental results show that the dynamic clustering results are more stable and more accurate by adding dynamic iteration to adjust the basis of cluster center. To a certain extent, the performance of network intrusion detection is improved, the feasibility and effectiveness of the algorithm are demonstrated.
出处
《现代电子技术》
2009年第20期85-87,共3页
Modern Electronics Technique
关键词
聚类
聚类中心
距离
迭代
入侵检测
clustering
cluster center
distance
iterative
intrusion detection
