基于数据挖掘的客户细分的建模与实现

华侨大学硕士学位论文基于数据挖掘的客户细分的建模与实现姓名：张永礼申请学位级别：硕士专业：管理科学与工程指导教师：彭霈20081201IIIAbstractThecustomerisoneofthemostimportantresourcesofanenterprise.Thecompetitionbetweenenterprisesfocusesoncustomers.Whetheranenterpriseholdsacustomerdependsonthestateoftherelationshipbetweentheenterpriseandthecustomer.Toimproveontherelationshipwiththecustomer,anenterprisemustcarryoutcustomerrelationshipmanagement(CRM).CustomeranalysisisthebasisofCRMandcustomersegmentationisanimportantitemofcustomeranalysis.Butthereisnotanyefficientcustomersegmentationmethodatpresent.Thepaperfirstlyintroducescustomerloyaltheory,DataMiningarithmeticandthetraditionalmethodsofcustomersegmentation,thenstudiesthecustomer-segmentationmethodaccordingtocustomerlifetimevalueandimprovesontheexistingcustomersegmentationmodel.Basedoncustomercurrentvalue,customerpotentialvalueandcustomerloyal,thenewcustomersegmentationmodeldividescustomersintoeightclasses,thenaccordingtothedifferentcustomers’class,theenterprisecantakedifferentmarketablestrategy.Accordingtothecustomer'spurchasehistoryrecords,weusetheassociationrulesalgorithmtopredictthefuturepurchaseandtheprobabilityofpurchase.Finally,wecangetthecustomerpotentialvalueaccordingasproductcost.Basedonthepresentstudystatusonthehomeandabroadresearch,weapplyclustering,decisiontreeandANN(ArtificialNeuralNetwork)algorithmtotheevaluationofcustomerloyalty.Finally,weappliedthecustomersegmentationmodeltooneenterprise,validatedit.AbstractIIIKeywords:CustomerSegmentation;Association;Clustering;DecisionTree;ANN(ArtificialNeuralNetwork)11.1(CustomerRelationshipManagement,CRM)CRM[1](CustomerSegmentation)[2]1)(CustomerPotentialValue,CPV)(CustomerCurrentValue,CCV)(CustomerPotentialValue,CPV)(CustomerPotentialValue,CPV)2)2KDD(KnowledgeDiscoveryinDatabase)AI()1.21.2.1SmithWendell1956[3]1)(Kolter)1995()()(CustomerDeliveredValue)(TotalCustomerValue)3(TotalCustomerCost)BarbaraJackson(1985)RobertsBerger(1989)BarbaraJackson(1994)BitranMondschein(1996)(RemainingLife)(LifeCycle)Courtheoux(1995)(CustomerLongtimeValue)()[4][5][6]2)WilkieCohen()SchiffnanHalev4[7][8]3)JinSellersArthurHughesRFM[9](Recency)(Frequency)(Monetary)20%20%60%(MigrateUp)RMFMarcus[10](CustomerLifetimeValue)[11][12]4)5[13]--[14]--[15][13]1.2.2KDD(KnowledgeDiscoveryinDatabase)1989811(The11thInternationalJointConferenceonAI)1999PAKDD158IEEEKnowledgeandDataEngineering1993KDDKDD6BCSinionFraserKDD(AmericanExpress)NSRC1%5O1%3587%SPN()X3:2(DatabaseMarketing)(CustomerSegmentation&Classification)(ProfileAnalysis)(CrossSelling)(ChurnAnalysis)7(CreditScoring)(FraudDetection)Web1)2)3)1.2.31)()CRM82)80/20(20%80%)5CRM3)4)(DecisionTree)5)1.31)92)3)CRMOLAP4)5)(CustomerCurrentValue,CCV)(CustomerPotentialValue,CPV)6)1.41)102)(CustomerCurrentValue,CCV)(CustomerPotentialValue,CPV)3)MicrosoftSQLServer2005SQLServerBusinessIntelligenceDevelopmentStudioSQLServer2005IntegrateServiceAnalysisServicesReportingServiceMDX()DMX()IntegrateServiceExcelAnalysisServicesOLAPOLAPReportingServiceWebExcelMicrosoftSQLServer20051.51)(CustomerLifetimeValue,CLV)11(CustomerCurrentValue,CCV)(CustomerPotentialValue,CPV)CLVCLVCLV2)1-11-1----------------3)(CustomerPotentialValue,CPV)(Association)(CCV)(CPV)121-11niijijjVprobprofit==×∑(1-1)iViijprobijijprofitijijj4)RMF[16]13(CustomerSegmentation)SuzanneDormer()2.12.1.1(Kolter)(1995)()(CustomerDeliveredValue)(TotalCustomerValue)(TotalCustomerCost)2-1142-11)2)3)4)155)6)7)[17]2.1.21)16[18]()()()(1)(2)(3)()[19]()172)(1)[20][21](2)[20]CRM[21]3)[19][21][22][23](1)18(2)(3)(4)()()()192.2FrederickReichheldRFM2.2.1RFMRFMJimSellersArthurHughesRFM[9]R(Recency)RRF(Frequency)M(Monetary)MRFM335554(1,1,1)(5,5,5),1252-2202-2RFM2003.8RFMR×F×M20%20%60%20%RFMRFMRFMRFM2.2.2RFM2-32-3RFMRFMMarcusRFM(2-3)21[10]2.3()()()2.3.1MartinChristopherAdrianPayneDavidBallantyneRelationshipMarketing[24]2-42-4()()222.3.2[25][26],2-52-5232.4[2]2.4.11)/CVCPCC=(2-1)CPCCCVCPm2-1242-150%n≥50%20%n≥20%10%n≥−n100%CPmnm−=×10%n−,2)1(1)TtttMQXNPVCi=−=−+∑(2-2)MtQtXCT/PVNPVC=3)25,10(1)TttttCPCICi=−−=+∑(2-3)tCPtCITC2.4.21)ABCABC2-62-6ABCA90%100%80%C90%20%100%B2680%20%70%20%10%2-62-6A,B,CABCABCA2)(1)(2-2)2-2ABCDE27(2)2-7(3)2-828----(4)[18]2-92-9292-32-32.52.5.1207065%85%[27][28](JeffreyGitomer)[29]1996(FrederickF.Reichheld)[30]302090FrederickF.ReichheldW.EarlSasserJr.[30]AlanS.DickKunalBasu(RelativeAttitude)[31]TuckerHallowellGremlerBrown[32]RichardL.0liver[33][34]2.5.2AlanS.DickKunalBasu19941)31[35]2)[35](1994)(1997)[36]3)-(1969)[31]S.J.BackmanJ.L.Crompton[37]2-10322-10BalogluCornelld[38]SPSS333.1(DataMining)3.1.1W.H.Inmon343-13-13.1.2OLAPOLAP(OnlineAnalyticalProcessing)OLTP(OnlineTransactionProcessing)OLTPOLAPOLTPOLAP1)OLTPOLAP2)OLTPOLAP3)OLTPE-ROLAPCube354)OLTPOLAPOLAP5)OLTPOLAP6)OLTPOLAP(BusinessIntelligent,BI)OLAP(Verification-DrivenDataAnalysis)(Discovery-DrivenDataMining)OLAPOLAP3-23-2BI3(OLAP)36DMOLAPDMOLAPOLAPDMOLAPDMOLAPOLAPDM[39]3.23.2.11)2)()3)374)X=YXY()()5)6)7)383.2.2API3-33-33.33.3.11)IBM50-80%API39DorlanPyleDataPreparationforDataMing(1)(2)(3)2)3)404)5)3.3.2()2SPSS5ACRISP-DM[39]1)5A5A5AssessAccessAnalyzeActAut