基于径向基函数的入侵检测系统  

Radial basic function-based intrusion detection systems

作　　者：李战春[1] 李之棠[1] 黎耀[1] 

机构地区：[1]华中科技大学计算机科学与技术学院,湖北武汉430074

出　　处：《计算机应用》2006年第5期1075-1076,1080,共3页journal of Computer Applications

基　　金：国家网络与信息安全保障可持续发展计划资助项目(2004研1917C021);湖北省自然科学基金资助项目(2004ABA051)

摘　　要：入侵检测系统(IntrusionDetectionSystem,IDS)是信息安全管理的重要组成部分,通过监测网络流量模式来检测入侵行为。在入侵检测系统中引入径向基函数神经网络,设计了一个基于径向基函数的网络入侵检测系统(RadialBasicFunctionsIntrusionDetectionSystem,RBFIDS)。该系统首先采集网络运行数据,然后采用K均值聚类算法确定RBF神经网络的系统参数。通过KDD99数据集对该系统进行性能测试,总的检测率达到98%,误报率为1.6%,表明该系统有较高的检测率和低的误报率。An intrusion detection system is a critical component for secure information management, and an intrusion detection system is an art to detect network intrusions by monitoring the network traffic patterns. In this paper, radial basic functions neural network was introduced into intrusion detection system and a network intrusion detection system based on radial basic functions neural network （RBFIDS） was designed. The system first gathers information of the network, then uses K-means clustering methods to determine the parameters of RBF neural network. The experiments were run on the KDD-99 datasets for performance evaluation, the system achieved a correct detection rate equal to 98% and a false detection rate equal to 1.6%. The experimental results show the RBFIDS system improves the performance of intrusion detection systems with a high detecting rate and a low false positive.

关 键 词：入侵检测系统 神经网络 径向基函数(RBF) 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]