1、中國科學技術大學碩士學位論文隱Markov模型和數(shù)據(jù)融合在入侵檢測中的應用研究姓名：安景琦申請學位級別：碩士專業(yè)：計算機應用指導教師：劉貴全20050501中國科學投術人學順：論文隱Markov模型發(fā)數(shù)拋融臺掃：入侵椅測中的應用研究AbstractTheopenspecialityofInternetprovidesthegreatconvenienceofsharingandinteractionofinformationMeanwh

2、ile，itraisesthegreatchallengeofsecurityofinformationSecurityofinformationhasbeenthekeyproblemofinformationsystemIntrusiondetectionsystem(IDS)，asapositivemeasureforinformationsecurity，hasmadeupthelimitationoftraditionalse

3、curityprotectionmeasureThroughbuildingdynamicsecuritycycle，itCallpromotetileprotectioncapacityofsystemandreducetilesecuritythreatsasgreataspossibleNowadays，SOmanynewattacksemergeendlessly，thatthetraditionalintrusiondetec

4、tionmethodscannotmeetthesystemsecurityrequirementsTheanomalydetectionteclmiquewhichcandetecttmknownintrusionbehaviorhasbecomeakeyresearchfieldinintrusiondetectionMoreoverwiththerapidlyexpandingofnetworkscaleandthegradual

5、lycomplicatingofthenetworkstructure，thedistributedintrusiondetectiontechniqueswhichcanbeappliedinlargescalenetworkbecomethepresentresearchkeyfield，tooAmongthesestudies，MultisensorDataFusionhasbeenfocusasanewresearchhotsp

6、otToapplythesenewtechniquesinactualsystem，thethesis，supportedbytheNationalScienceFundKeyResearchProject，‘GreatScaleNetworkOrientedDistributedIntrusionDetectionandAlertModel’，basedonacommercialIDSproduct——’’KezhenWangzhen

7、”networkIntrusionDetectionSystemasaresearchingfoundation，developresearchworksTilemajortasksandinnovationsasfollow：1)Anomalydetectioncanfindunknownintrusions，butitusuallyrequiressufficientsecuritybackgroundknowledgeItisve

8、rysignificanttoidentifysystemanomalybehaviorunderthespecialconditionofpoordomainknowledgeInthispaper，Markovchainmodelusedforanomalydetectionisdeeplydiscussedfromthreeaspects：one—stepmultistepMarkovchainandmulti—stepMarko

9、vchain—basedsequencepredictionTheexperimentsshowthemethodcandetecttheanomalybehaviorofsystem，whichdoesnot、necessarilyneedanysystemsecurityknowledge2)ComparedwithMarkgovModel，HiddenMarkovModel(HMM)cansimulatemorecomplicat