搜索
:2005-03-07;:2005-05-14:863(2002AA135230-1);(40471111):(1972-),,,,GISE-mail:wanghq@lreis.ac.cn1,2,3,1(11,100101;21(),257061;31,100039):,GIS,:;;;;GIS:P208:A:1672-0504(2005)04-0006-05,GIS,[1],,,[2];,,,,(SpatialDataMining,SDM),,()(,)()3[2-4]1111,,(,),[5-7]:1),,2),,3),4),,,,5),6),21420057GeographyandGeo-InformationScienceVol.21No.4July2005112GIS,[7-12],,(1)(generalgeometricknowl2edge),,(2)(spatialdistributionregulari2ties),,;,;(3)(spatialassociationrules),,(4)(spatialcharacteristicrules),,,(5)(spatialdiscriminaterules),(6)/(spatialclassificationrules/spatialregressionrules):,,;,,,,,,(7)/(spatialclusteringrules/spatialfunctionaldependencyrules):,,,,,,(8)(spatialserialrules),,,(9)(spatialevolutionrules),,GISGIS(snapshot),(10)(objectorientedknowl2edge)2211,HanHolsheimerMatheus,Matheus[7,11]Matheus,,,,DB4DB,,(R);,;,;,4,212:1),,2)74:3),,4),5),,6),,,7),3,Web;()()(BP);[2,3,11-18](1)(statisticalanalysismethod),,,,,,,,(2)/(induction/generalizationlearningmethod),,QuinlanC510,,,,,(3)(methodsusingclustering)/,,4:K-K-EM,,,BirchCureChameleon:,,,DBscanOpticsDenclue,,StingSting+WaveClusterClique(4)(methodsusingclassification):RSGIS,,;,:,;,(5)(ExploratoryDataAnalysis,EDA),EDA,(ExploratorySpatialDataAnal2ysis,ESDA)[19]EDAESDA,,(6)(spatialtrendanalysis),,,821,,,(y=X+),SAR[19]:y=Wy+X+(7)(),,[20]:(decisiontree)(artificialneuralnetwork)(8),,,,,,,,,,,4GIS(GIS)3,,GIS,GISGISGIS,3[21-23]:1)GIS,2)GIS,GIS,GIS,,GISGIS:GISOLE(),Geomin2er;GIS,GIS,GIS3)GISGISOLE,,GIS5,:,,,,,,,,[2,3]SDM,,SDM,SDMGIS[11,12]:(1)GB,TB,,,[15],,,(2),,,,,,(3),,,(4)Web,,()94:(5),,:[1],,.[M].:,2000.3-10.[2],,,.[J].(),2001,26(6):491-499.[3],,,.[J].(),2002,27(3):221-233.[4]HANJ,KAMBERM.DataMining:ConceptsandTechniques[M].SanFrancisco:AcademicPress,2001.145-176.[5],,,.[J].,2001,6A(9):854-860.[6],.[J].,2003,19(3):32-34.[7]MILLERHJ,HANJ.Geographicdataminingandknowledgedis2covery:Anoverview[A].MILLERHJ,HANJ.GeographicDataMiningandKnowledgeDiscovery[C].London:TaylorandFran2cis,2001.3-32.[8],,.[J].,2002,2:11-13.[9],,,.[J].,2003,20(2):132-134.[10],,.[J].,2004,20(2):38-40.[11]ESTERM,FROMMELTA,KRIEGELHP,etal.Spatialdatamining:databaseprimitives,algorithmsandefficientDBMSsup2port[J].DataMiningandKnowledgeDiscovery,2000,4(2-3):193-216.[12]GAHEGANM,WACHOWICZM,HARROWERM,etal.Theintegrationofgeographicvisualizationwithknowledgediscoveryindatabasesandgeocomputation[J].CartographyandGeographicInformationSystems,2001,28:29-44.[13],.[J].,2003,29(6):9-10.[14],.[J].,2002,17(4):198-204.[15]AGGARWALC,HANJ,WANGJ,etal.Aframeworkforpro2jectedclusteringofhighdimensionaldatastreams[A].FREY2TAGJC,LOCKEMANNPC,ABITEBOULS,etal.Proceed2ings2004VLDBConference:The30thInternationalConferenceonVeryLargeDatabases(VLDB)[C].SanFrancisco:MorganKaufmann,2004.852-863.[16]DONGG,HANJ,LAMJMW,etal.Miningmulti-dimensionalconstrainedgradientsindatacubes[A].APERSPMG,ATZENIP,CERIS,etal.VLDB01:Proceedingsofthe27thInternationalConferenceonVeryLargeDataBases[C].SanFrancisco:MorganKaufmann,2001.321-330.[17]SHEKHARS,LUC,ZHANGP.Aunifiedapproachtospatialoutliers[J].GeoInformatica,2003,7(2):139-166.[18],.[J].,2003,3:5-7.[19]HAININGR,WISES,MAJ.Exploratoryspatialdataanalysisinageographicinformationsystemenvironment[J].TheStatisti2cian,1998,47:457-469.[20],,.[J].,2004,20(3):89-91.[21],,.[J].,2002,31(4):304-309.[22],.GIS[J].,2003(2):21-24.[23]GOODCHILDMF,HAININGRP,WISES,etal.IntegratingGISandspatialanalysis:problemsandpossibilities[J].Interna2tionalJournalofGeographicalInformationSystems,1992,6(5):407-423.ResearchonProgressofSpatialDataMiningWANGHai-qi1,2,3,WANGJin-feng1(1.LREIS,InstituteofGeographicalSciencesandNationalResourcesResearch,CAS,Beijing100101;2.CollegeofGeo-ResourcesandInformation,UniversityofPetroleum(EastChina),Dongying257061;3.GraduateSchoolofChineseAcademyofSciences,Beijing100039,China)Abstract:Spatialdataminingdiscoversandextractsimplicitknowledge,spatialrelations,spatialrulesandotherinterestingpatternsfromspatialdatawhichhavecomplexitycharactersofmass,nonlinearity,multi-scale,high-dimension,fuzzyetc.Thispaperat2temptsanannotatedreviewofknowledgetype,methods,architecture,processandintegrationwithGISwhicharerelatedwiththefieldofspatialdatamining,andespeciallypaysattentiontoknowledgetypeincludingspatialcharacteristicanddiscriminaterules,spatialclassificationandclusteringrules,spatialdistributionandassociationrules,spatialserialandevolutionrules,andatthesametime,paysattentiontodataminingmethodsincludingstatisticalanalysis,machinelearning,exploratorydataanalysis,visualizationanalysis,etc.Inaddition,theexistingproblemsandresearchdirectionsofspatialdataminingarealsoanalyzedaboutitstheory,applicationandimple2mentofsoftwaresystem.Inconclusion,themaintrendsofspatialdataminingareexpectedthroughseamlessintegrationwithdatabase,knowledgedatabase,expertsystem,decisionsystem,visualization,andnetwork.Keywords:spatialdata;knowledgetype;datamining;architecture;GIS0121