欢迎来到天天文库
浏览记录
ID:36570868
大小:3.00 MB
页数:74页
时间:2019-05-12
《基于数据挖掘的入侵检测方法及系统研究》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、中国农业大学博士学位论文基于数据挖掘的入侵检测方法及系统研究姓名:齐建东申请学位级别:博士专业:农业电气化与自动化指导教师:陶兰20030601中国农渡丈学簿士学垃论文AbstractIntrusiondetectionsystemqDS)playsimportantrolesintheinformationsecurityarchitecture。Thecomputercriminalismoreandmorepressinganddangerousnowadays,whichposesurgentdemandsontheperf
2、ormanceofID&ThebigshortcomingofculTentIDSisunabletodetectintrusionbehaviorquicklywhenfacinglargeamountofauditdata,unabletodetectnewtypeofattacks,andhighfalsepositiveratewhichinfluencegreatlytheperformanceof/DS.Anewintrusiondetectionapproachwasputforwardinthisdissertati
3、onandaNetwork-basedAnomalyIntrusionDetectionSystem(NAIDS)wasimplementedbasedonthoseapproaches.Basedonthedetailedandcomprehensivestudyondataminingbasedintrusiondetectiontechniques,NAIDSapplytheassociationroleandclassificationtechniquesintodetectingintrusionbehavioramong
4、networkauditrecordfromanewperspective。Aspecttoassociationrulesmining,constructingtwominingmodes:staticmininganddynamicmining;implementingtwolevelmining:single*levelmininganddomain~levelmining.Aboutclassificationengineering,themainstreamclassificationtechniqueswelgcompa
5、redthroughthoroughlyexperiments,andsomeimprovementwasmadetodecisiontreetowardtheconcreteproblem,whichmakeNAIDSdetectsomenewtypeattacksandthiskindofcapabilityembodiestheadvantageofanomalydetectionovermisusedetection;incrementalminingapproachwasputforwardwhichdetectonewi
6、ndowdataamount,insteadofbatchoftc∥iprecord,whichwasverysuitabletoon—lineminingandmakeNAIDSbehilghreal-timeperformance.Researchworkondataminingbasedintrusiondetectionapproacheswhichhasbeendonebelongstothefieldofmisusedetectioninnature,associationrulesandfrequentepisodes
7、miningaimtodescribetheintrusionsignature.therulerclassifierwasusedtOmainlydetectintrusionbehaviorNAIDSisthefimtdataminingbasedanomalydetectionsystem,thefirstintrusiondetectionsystemwhichlowerfalsepositiveratebyclassificationengineering,thefirstintrusiondetectionsystemw
8、hichputforwardslidingwindowstechniquestocarryoutincremental,on-linemining。Inprincipal,dynamicslidingwindowmakeNAIDSha
此文档下载收益归作者所有