PSO算法和神经网络的入侵检测系统设计.pdf
文本预览下载声明
文章编号:1671—4598(ZOLO}08—1924—04中圈分类号:TP311 文献标识码:A
PSO算法和神经网络的入侵检测系统设计
龚 娟,段树华
(湖南铁道职业技术学院,湖南株洲412000)
摘赛:针对入侵检测系统检测率低,整体性能不好的问题,在探讨入侵检测技术和人工神经网络理论的基础上,提出了一种基于
PSO算法优化的径向基函数神经网络的入侵检测系统.采用具有全局寻优的功能PSO算法,该算法能够改进传统的RBF神经网络学习
策略,弥补RBF神经网络参数设置的不足,采用了来自KDDCUP99的权威数据来进行网络学习和测试,在此基础之上,进行了入侵检
测系统的设计与实现,实验结果表明,基于PSO和RBF神经网络的人侵检测系统有效地提高了入侵检测的效率.
关键词:PSO算法;入侵检测系统;人工神经网络
ofNeuralNetworkBasedonParticle
Application
Swarm forIntrusionDetection
Algorithm
GongJuan,DuanShuhua
(HunanProfessional 412000,China)
Rilway TechnologyCollege,Zhuzhou
Abstract:AimtotheIOW ofIntrusionDetectionandthe ofoverallissues,it theintrusionde—
efficiency System poorperformance explore
teetion andthe ofartificialneuralnetworkbasedonPSO basedon ofradialbasisfunctionneuralnet—
technologytheory algorithmoptimization
workintrusiondetection withthefunctionof PSO can thetraditionalRBFneuralnet—
system.using globaloptimizationalgorithmimprove
theKDD
work neuralnetworkcanmakeforthelackof theauthorityfrom CUP99datafore--
learningstrategy,RBF up parameters,using
resultsshowthat
and thebasisofon ofaIntrusionDetectionDesignand the
learningtesting,on top System Implementation,experimental
the
PSO—basedRBFneuralnetworkandintrusiondetectioneffectivelyimprove ofintrusiondetection.
system efficiency
words:PSOalgorithm;intrusiondetection;ststem;ANN
Key
0 引言
显示全部