欢迎来到天天文库
浏览记录
ID:8365074
大小:355.00 KB
页数:40页
时间:2018-03-22
《problem statement》由会员上传分享,免费在线阅读,更多相关内容在教育资源-天天文库。
1、DataMining40DataMiningApproachesforNetworkIntrusionDetectionFinalReport11/9/2006Group9KarlaBracamonteJeffreyGawlinskiJordanHarstadOmarRodriguezMichaelWrightDataMining40AbstractOverthepastyears,muchinteresthasbeenshowsindataminingtodetectnetworkintrusions.Thispaperprovides
2、methodologyandthoughprocesswithinthediscussedtask.Baseduponexperiencesinsuchafield,dataminingtechniquesarefurthersuggestedandwillvaryuponexpertiseandnetworkinfrastructure.Thispaperisintendedfortheusebycomputerandnetworksecurityprofessionalswhowishtostudyandlearnmoreaboutt
3、hescienceofdatamining,aswellasresearchwaysexpertsusecurrentdataminingmaterialaswaystofurthertheirknowledgeinintrusiondetection.DataMining40ProblemStatementThegoalofIntrusionDetectionSystems(IDS)istodetectanintrusionasithappensandbeabletorespondtoit.Aprimaryconcernwhenimpl
4、ementingIDS,especiallyintheformofDataMiningistheresultof“falsepositives.”Afalsepositivecouldbeasituationwheresomethingabnormaloccurs,butisnotnecessarilyanintrusion.AhighpercentageoffalsepositivesmayrenderanIDSuselessandcauseuserrevolts.Anotherproblemisthatof“falsenegative
5、s”,inwhichanintrusionisactuallyoccurringbutitgoesundetectedbyIDSorIDSmisclassifiestheevent.ChallengesandMotivationConsequently,challengesmaypotentiallyarisewithinallintrusiondetectionmethods,especiallythatofdatamining.Thebroaderfieldofthesechallengesisreferredtoas“threata
6、nalysis”andthedefinitionofwhatitentails.Thus,theunderlyingneedthatmotivatessuchresearchistoprovidereliabilitytothosecompaniesandorganizationsthatdependondata.Bestdefined,threatanalysisisthestudyofknownandunknownpatterns,whichanalyzecertaindatatrafficpathsandtheirfollowing
7、characteristics.Statisticalandinformationstateresearchonanythingfromneighborhoodsandtheirdemographicscanbealsoincludedwithinthreatanalysis,thusmakingthecharacterizationprocessforintrusiondetectionquitebroadanddifficult.Oneofthebestmethodsinprovidingreliabilityofalldatatha
8、tistrackedistoreducefalsepositivesanderrors,whileincreasingconsumerconfidenceindatamanagementpro
此文档下载收益归作者所有