数据挖掘技术在入侵检测中的应用  

APPLICATION OF DATA MINING IN INTRUSION DETECTION

作　　者：李云[1] 刘学诚[1] 朱峰[1] 

机构地区：[1]泰山学院数学与系统科学学院,山东泰安271021

出　　处：《计算机应用与软件》2011年第5期117-119,共3页Computer Applications and Software

基　　金：山东省教育厅科技发展项目(J07WJ23)

摘　　要：针对入侵检测系统的研究现状和面临的问题,研究了数据挖掘技术应用到入侵检测中的优势,分析了当前基于数据挖掘的入侵检测中存在的不足。针对目前基于数据挖掘的入侵检测时空效率不高的问题,对频繁模式算法进行了研究,改进了频繁模式算法,用两步模式增长代替一步模式增长模式来加快挖掘速度,并且增加时间特性、属性相关和轴属性加以约束。通过试验证明改进后的算法在时空效率上得到了改善,减少了扫描数据库的时间和生成无意义的模式,提高了规则的有用性。In this paper we analyse the research status quo of intrusion detection system（IDS） and the problems the IDS is encountering,study the advantage of applying data mining technique to IDS,and analyse the deficiency of existing data mining-based IDS.For the problem of the time and space inefficiencies in data mining-based intrusion detection,after studying the frequent pattern algorithm,we improve this algorithm by using two-step growth model instead of one-step model to boost the mining speed,and constrain it by adding time feature,attribute correlation and axis attribute to it.Examination proves the amelioration of the improved algorithm in time and space efficiency,and the time of database scanning as well as the generation of less meaningless pattern are all decreased,thus the availability of rule is improved.

关 键 词：网络安全 入侵检测 频繁序列 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]