1种基于可变精度粗糙集的网络入侵检测模型

A NETWORK INTRUSION DETECTION SYSTEM MODEL BASED ON VARIABLE PRECISION ROUGH SET

当前各网络入侵检测算法的准确率仍不尽人意;针对此问题,提出1种基于可变精度粗糙集(Variab le Prec isionRough Set)的网络入侵检测模型。模型通过粗糙集对不确性数据进行筛选,再利用粒子群算法对数据进行约简,然后再根据设定的阀值,用可变精度粗糙集导出规则并得到检测结果。实验结果表明,本模型运用的粒子群算法数据约简速度高于利用遗传算法的同类模型,且基于可变精度粗糙集的入侵系统检测准确率比基于非可变精度粗糙集的检测系统高。

With uncertain measurements of rough set, relative attribute reduction is done by applying PSO （particle swarm optimizer）. Then VPRS is used to deduce the attributes and generate roles under the threshold value. Experiment results show that this model has a greater reducing speed and a higher accuracy than models that use GA.