聚类分析方法及工具应用研究

2006Vol.33.2(250014),(SPSSDBMiner),,,155,I)Bine:ResearchofClusteringMethsdlO15inAPPlicationWANGXinWANGHOllgGuoZHANGJian-xiGUJianesjunHJEoesFang(InformationManagementSchoolofShandongNormalUniversity,Jinan250014)AtractClusteri15animrtantresearehdireetioninthefieldoftaMining.ThisPaperintrodueestheeoneeptandmainmethodsofelustering,analyzesandeomparestheelusterifunetionsofoverseasleadingdataminitools(SPSSandDBMiner)basedontherealexamples.Theeonelusion15giveninthelastofper.KeywordsDatamini,Clusteri,Clusteritool,SPSS,DBMiner1,(DataMining),(Kl)l)),,,(clus-ter),,1.1,,,,,,,,Web,,,,1.2:,,,2:(1):KMEANS,KME-11S,CLARANS;(2):BIRCH,CIJRE,CHAMEIE()N;(3)DlAN,OP-11CS,DENCLUE;(4):STING,CLIQUE,wAVE~CLUSTER;5):1.3:SASEnterpriseMinerIBMIntelligentMinerHMinersetSPSSClementineSimonFraserDBMinerRuleQuestResearehSees,CoverstoryEXPL()RAKnowledgeDiseoveryWorkbenehQuit2SPSSClementine2.151CleITntineSPSS6(StatistieSPaekageforSoeialScienee),SPSSClementine,,,eRlsRDMelementine,,),,SP,SPSSforWindowsVn.,2.2SPSS,:(Hierar-ehiealClusterProeedure)(IterativePartitioning,:,,,197Procedure),,,(agglomerative)(divisive),,,,,,:(1):;(2);(3);(4)23,,,,,2.3512.3.1,2.3.2,:(l)();(2);(3);(4)(q);(5)(),2.3.3(l):1(1)(2),,:a)(),b)(ST)(),,,(3),,2.3.4(1)()()(2)()dtj,()ij,d,(),,],:,t;i,],t,(),(3)(2),()(2)k,(2),k+12.3.5,:1:2:3:4:,,,,,198,,,SPSS2.351,11,,ases(),Between-grouPshnkage(,),reEuelideandistanee(),,SPSS:DendrograusingLverageLinkage(BetweenGrouPs)CASEbe1NumResealDistanceClusterCombine0510152025~---~----97l236l15I2132””,.xx”rŽ””KƒŽXK”x3Ž,Jx“xŽx””xxKŽŽŽ2,\,(2):(),Procedure),,,(agglomerative)(divisive),,,,,,:(1):;(2);(3);(4)23,,,,,2.3512.3.1,2.3.2,:(l)();(2);(3);(4)(q);(5)(),2.3.3(l):1(1)(2),,:a)(),b)(ST)(),,,(3),,2.3.4(1)()()(2)()dtj,()ij,d,(),,],:,t;i,],t,(),(3)(2),()(2)k,(2),k+12.3.5,:1:2:3:4:,,,,,198,,,SPSS2.351,11,,ases(),Between-grouPshnkage(,),reEuelideandistanee(),,SPSS:DendrograusingLverageLinkage(BetweenGrouPs)CASEbe1NumResealDistanceClusterCombine0510152025~---~----97l236l15I2132””,.xx”rŽ””KƒŽXK”x3Ž,Jx“xŽx””xxKŽŽŽ2,\,(2):(),6Manager(5):,(6)(7)KMeanS,KMeanS.,,,,,,,(2):,(3):,(4):,,SQLSever(5):ClementineSolutionPublisherClementineCSQI,(6),(),,4.2DBMiner(1);;DMQI,(2):ex-cel(3):,(4):MierosoftSQI,Sever()LAIHanJiawei,KarnberM.I)ataMlning:ConeeptsandTeehniqu[(j.MonganKaufmannpublishors,2000.225278AlsabtiK,RankaS.SinghV.Anefieientmeanselusteringalrithm[A.In:IPPS98,Proe.otheFirstWorkshoponHighPe:fornaneeDateMlning(.()rlando,Flor:da,USA,1998EterM,KrlegelHP.SanderJ,eral.ADensiry-BasedAlgorithm,.forDseoveringClusterslnIargeSpatialDatabaseswithNoie.In:Proe.ZndInternationalConfereneeonKnowledgenseoveryclDataMiningC.Portland,()R.1996.226231wanolxx.zanioloc.tabasesvsremExtension,foreisionlSu000rr:theAxlAolr,aeh[A.In:ACMSIGM()DWorkshooonlResearehIssueSinDataMiningandKnowledgeseoveryLC,2000.11~20...23(9)::1~3.:SPSS[M.:,2002ZllangT,etal.lllRCHAneffieientdataelusteringmethodforverylargedatabases.In:Proe.oftheACMSIGM()DInt1CononManagementofDal:2.Montreal:ACMPress.1996.73~84,..,2(4):42~45(190),Max-CodeFMCSG(FindingMaTnalCompleteSubgraPh),Max--de,,FMC.,FMCSCWashioT.KunioN,etal.CompleteMiningofFrequenrpatlernsfromGraphs:MiningGraphData.MaehzngIearxling,2003,50(3):321~354SynlthP,GoodmanRM.AnInformationTheoretieApproaehtoRuleInduetlonfromDatabase.IEEETrans.KnowledgeDataEng.,1992,4(4):301~316ChenAL,TangCJ,etal.improvedalgorithrnbasedonrnaxi-mL.meliqueandFPtreeforminngassoeiationrules.JournalofSoftware,2004,15(8):11981207IJoukakisE,Tsouros(.Adepthfirtsearehalgorirhmtogeneratethe[amilyofmaximalindndontsetsofagraPhlexzeographi-ly.omPuting,1981,27:249266KopfR,RuhcG.Aeor-nputazionalstudyoftlewghtedindepend*entsetproblemforgeneralgraphs.FoundationsofControlEngi.neering,1987.12(4):167~18()BonizeIM,I,elilloM.StlxV.AroxmaringthemaximumweigItelquoungrololeotordynarnie.IEEETrans.NeuralNet7work.2000,11(6):12281241BlggN.Algolra:egra)ltheory.CambridgeUnvorityPress,Canbridgc,England.1974Adleman1M.Moleeulareomputationofsolutiontoeombinatoalo)timization.Scienee.1994,226:10211024Sun51.AlgebraSrruerure.ChinaUniversityofScieneeanTeehnologyPress.Hefei.hina.199o200