资源描述：
《NNCluster An Efficient ClusteringAlgorithm for Road Network Trajectories版本》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、NNCluster:AnEfficientClusteringAlgorithmforRoadNetworkTrajectoriesGook-PilRohandSeung-wonHwangDepartmentofComputerScience&Engineering,PohangUniversityofScienceandTechnology(POSTECH),Pohang,RepublicofKorea{noh9pil,swhwang}@postech.ac.krAbstract.Withtheadventofubiquitouscomputing,weca
2、neasilyac-quirethelocationsofmovingobjects.Thispaperstudiesclusteringprob-lemsfortrajectorydatathatisconstrainedbytheroadnetwork.Whilemanytrajectoryclusteringalgorithmshavebeenproposed,theydonotconsiderthespatialproximityofobjectsacrosstheroadnetwork.Forthiskindofdata,weproposeanew
3、distancemeasurethatreflectsthespatialproximityofvehicletrajectoriesontheroadnetwork,andanefficientclusteringmethodthatreducesthenumberofdistancecompu-tationsduringtheclusteringprocess.Experimentalresultsdemonstratethatourproposedmethodcorrectlyidentifiesclustersusingreal-lifetra-jector
4、ydatayetreducesthedistancecomputationsbyupto80%againstthebaselinealgorithm.1IntroductionDuetotheevolutionofpositioningandsensortechnologiessuchasGPSandRFID,wecannoweasilyrecordthemovementsortrajectoriesofmovingobjects.Someexamplesofthisincludevehiclelocations,objecttrackingdata,and
5、animalmovementdata.Recently,amassiveamountofcollectedtrajectoryinformationhasbeenpub-lishedandsharedinwebsitesorinwebcommunities[1,2,3].Whilethesesitessimplyvisualizetherelevanttrajectories,advanceddataanalysistechniques,suchasclustering,canbeusedforrecommendinginterestingtravelseq
6、uencesbasedonthecommontrajectoriesofusers[28]orfindinguserswithsimilarlifeexperiencesbasedontheirtrajectories[22].Thispaperstudieshowtodevelopanefficientclusteringalgorithmfortra-jectorydata.Clusteringtrajectoriescanbeusedtoidentifydistinctgroupsinwhichtrajectorieshavemoresimilarmovin
7、gpatternsthanthoseinothergroups.Specifically,thispaperfocusesonclusteringthevehicletrajectoriesofusersonThisworkwassupportedbyMicrosoftResearchAsiaandEngineeringResearchCenterofExcellenceProgramofKoreaMinistryofEducation,ScienceandTechnol-ogy(MEST)/KoreaScienceandEngineeringFoundat
8、ion(KOSEF),grantnumberR11-
