时空k-匿名集数据的关联规则和序列模式挖掘分析-association rules and sequential pattern mining analysis of spatiotemporal k - anonymous set data.docxVIP

AbstractWiththedevelopmentofLocation-BasedServicessystem(LBS),peoplearegoingtopaymoreandmoreattentiontoprivacyprotectionofLBS.Then,privacyprotectionofLBSisbecomingahotspottopicinGISandmobilecommunicationdomains.AlthoughaseriesofLBSprivacyprotectionmethodshavebeenproposed,theycan’tresistattacksbasedondataminingtechniques.Inordertoverfythisproblem,thispaperconductedthefollowingresearch:(1)basedontheanalysisoftherelatedliteraturetheory,firstproposedthetraditionalspatio-temporalK-anonymousanditsvariantsmaycannoteffectivelydefenseanalysisinferenceattackproblemoflargespatialandtemporalscalesbasedonanonymousset,forthispurpose,thedesignanddevelopmentofthesoftwaresystemofLBSspace-timeK-anonymityexperimentaldatageneratedfromGPStrackdatasimulation.TheresearchobjectofthisprojectisaseriesofanonymousdatageneratedinLBSspace-timeK-anonymousqueryprocess.ButatpresentSpatialTemporalK-anonymousandoptimizationmethodhasnotbeenwidelyusedinthecommercialLBSsystem,sowefirstlypreprocesstheGPStrajectorydata,theexperimentalsimulationofanonymitysetdatatoacertainscalegeneratedfromGPSdatabasedonspatio-temporalK-anonymousalgorithmproposedbyGruteseretal.GPStrajectorydataofthisexperiment,fromataximanagementcompanyinNanjingcityinJuly15,2007,collection(0：00-24：00)2612taxiGPStrajectorydataperiod.Experimentalsimulationofanonymousdataproductioninclude：LBSsnapshotqueryanonymityset,LBScontinuousqueryanonymityset.(2)wedesignedthemethodsofminingassociationrulesandsequentialpatternsfromLBSSpatialTemporalK-Anonymitydatasets,thenusethetwomethods,withanonymitysetsnapshotdatageneratedfromasimulatedGPSdataandtheanonymitysetsequenceasexperimentaldata,respectivelyfortheminingexperimentanonymitysetassociationrulesandsequentialpatterns.Theexperimentalresultsshowthat,Concerningthesensitiveregionsofspace-timeprivacyinferenceattack,basedonanonymousdataassociationrulesminingmethodisproposedinthispaperandthesequencemode,resultinginmorethreateningprivacyrisk.(3)firstproposedbydynamicanalysisoftheassociationruleprivacysensiti