计算机工程与应用 ›› 2010, Vol. 46 ›› Issue (21): 82-85.DOI: 10.3778/j.issn.1002-8331.2010.21.023

• 网络、通信、安全 • 上一篇    下一篇

基于簇连接度聚类算法的入侵检测

党小超1,郝占军2,3,王筱娟3   

  1. 1.西北师范大学 网络教育学院,兰州 730070
    2.西京学院 工程技术系,西安 710123
    3.西北师范大学 数学与信息科学学院,兰州 730070
  • 收稿日期:2010-03-26 修回日期:2010-05-24 出版日期:2010-07-21 发布日期:2010-07-21
  • 通讯作者: 党小超

Intrusion detection based on cluster connectivity clustering algorithm

DANG Xiao-chao1,HAO Zhan-jun2,3,WANG Xiao-juan3   

  1. 1.College of Network Education,Northwest Normal University,Lanzhou 730070,China
    2.Department of Engineering Technology,Xijing University,Xi’an 710123,China
    3.College of Mathematics & Information Science,Northwest Normal University,Lanzhou 730070,China
  • Received:2010-03-26 Revised:2010-05-24 Online:2010-07-21 Published:2010-07-21
  • Contact: DANG Xiao-chao

摘要: 针对K-means、FMC聚类算法容易陷入局部最优且对初始解很敏感的问题,提出了一种新的基于划分和连接度的聚类优化算法,并给出了具体算法实现,明显地避免了对初始化选值敏感性的问题。给出了在KDDCUP99数据集上的检测结果,实验表明该算法具有较高的检测率及较低的误检率。

关键词: 入侵检测, 聚类算法, 连接度,

Abstract: In view of the problem that K-means and FMC cluster algorithm easily trap in a local optimum and strongly depend on the initialization,a new clustering optimization algorithm is presented based on division and connectivity,and gives concrete realization of the algorithm,avoids the sensitivity to the initialization value obviously.The experimental results based on the datum of KDDCUP99 demonstrate that the algorithm possesses higher detection rate and lower misuse detection rate.

Key words: intrusion detection, clustering algorithm, connectivity, cluster

中图分类号: