改进的蚂蚁聚类入侵检测方法

Improved Ant Clustering Intrusion Detection Method

在现有的自适应蚂蚁聚类算法中,自适应参数的调整往往凭经验取值,从而影响聚类质量。针对该问题,提出一种利用快速模拟退火算法实现蚂蚁聚类自适应参数动态调整的改进方法。基于该算法构建的入侵检测系统无需预先指定簇的数目,也不要求满足正常行为的数目远大于入侵行为的数目等条件。对KDD CUP1999数据集的仿真实验结果表明,该算法可以得到较理想的聚类,对未知入侵有较好的检测效果。

In present adaptive ant clustering algorithms,the values of adaptive parameters are adjusted by experience,which affects clustering quality.To solve the problem,this paper proposes an improved method by using fast Simulated Annealing Algorithm（SAA） to realize dynamic adjustment of ant clustering adaptive parameters.An intrusion detection system based on the algorithm does not require pre-specification of the number of clusters,or that the number of normal behaviors is far greater than the number of intrusions.Simulation results on KDD CUP1999 dataset show that the algorithm can get better clusters,and has better detection effects on unknown intrusions.