基于数据挖掘的y某融资租赁公司客户细分分析

《基于数据挖掘的y某融资租赁公司客户细分分析》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、RESEARCHOFCUSTOMERSEGMENTATIONINTHEYDFINANCIALLEASINGCO.LTDBASEDONDATAMININGABSTRACTAtpresent,therecommendationsystemmainlyusesthetechnologyofcollaborativefiltering(CF).CFtechnologyisbasedontheassumptionthatallcustomersareofsamevaluetothecompany,usingthe

2、resultgeneratedbythistechnologymayhavethecompanyspendmoneyonthelessvaluablecustomersanddoesnotgetnicepayback.Recentlyanewtechnologybasedonthecustomerlifetimevaluegainshighvisibility,thekeyideaofthistechnologyistocategorizethecustomerusingtheweightedsumof

3、threevariablesfromthecustomerlifetimevalue.Withthesevariableswecanrecommendthecustomerwithmoreflexibilityandaccuracy.ThekeydrawbackofthismethodisthatitrequireenoughdatabeencollectedtogetaccurateresultsfromtheCLV,otherwisefaultyresultsmaybegeneratedfromth

4、erecommendationsystem.Inthispaper,wefirstgivebriefintroductionoftheaforementioned万方数据recommendationtechnology,includingthecalculationmechanismandexecutionprocess,andthencarefullyanalyzetheproandconsofthetwomethods.Basedontheinformationanewhybridrecommend

5、ationmethodhasbeenproposed,arealworldexperimentwiththeapplicationofthisnewmethodhasbeenconductedtoexaminethefeasibility,resultsdemonstratetheimprovementofthisnewmethodoverthepreviousones.Thekeyinnovativepartofthispaperistheintegrationoftheprosofthetwoexi

6、stingtechnology,theutilizationoftheconceptfrombothcustomerpreferencemethodandcustomerlifetimevaluemethodtobevitalpartofthenewlyproposedmethod.Themoreimportantistheintroductionofthethresholdconcept,sothecombinationofthetwomethodscanbefluentandadjustableto

7、avoidthedeviationofusingthetworecommendationmethodsseparately.Thekeyideaofthenewmethodisatwo-stepoptimization,firststepistoidentifytheloyaltylevelofthetargetcustomersusingCF,thisisusedtoclassifythecustomerswithareasonablethresholdandthesecondstepistoappl

8、ytheweightedRFMmethodtodifferenttypesofcustomerscategorizedbythethresholdgeneratedinthefirststep.Therecommendationsystemcanbeadjustedbasedonthepropertyofdifferentgroupofcustomertomaximizethereturnfromtheproductsrecommendat