基于数据挖掘技术的移动通信行业客户细分

41720077JOURNALOFSHANGHAIJIAOTONGUNIVERSITYVol.41No.7Jul.2007:2006208211:(19662),,,,.(Tel.):021252302558;E2mail:[e-mail].:100622467(2007)0721142204,,(,200052):K2Means,,..,,.:;;K2Means:F279.23:ATheResearchofCustomerSegmentationBasedonDataMininginMobileTelecommunicationIndustryZHOUYing,LBWei,JINGMiao(AntaiCollegeofEconomics&Management,ShanghaiJiaotongUniv.,Shanghai200052,China)Abstract:ThispaperbuiltacustomersegmentationmodelbytherulefunctionoferrorsquaresumofK2Meansclusteringfromdataminingtechnology,whichcanprovideascientificbasisforthemarketingstrat2egy.Itpresentsapositiveeffectonenterprisesbecauseofitspracticabilityandsimplicity.Keywords:datamining;customersegmentation;K2Meanscluster,,.20068,4,43747.5.,,ARPU..,,,,.,.,.,.1.,..William[1],AIO(ActivityInterestsOpinion)[2].,()©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.[3][4].Haley[5],.(CLP),[6].,,,,.,,,.,,.,,.,.,,.,,,.2,.K2Means,CRISP2DM(CrossIndustryStandardProcessforDataMining)(1).6,.1CRISP2DMFig.1StandardproceduresofCRISP2DMcustomersegmentationmodelbuild.,..,,,,.,,,.SAS8.0K2Means,:(1)n,I=1,KZj(I)j=1,2,,K(2)D(Xi,Zj(I)),i=1,2,,n,j=1,2,,K;D(Xi,Zj(I))=min{D(Xi,Zj(I))}XjWK.(3)I=I+1,Zj(I+1)=1njnjiXjij=1,2,,KJc(I+1)=kj=1njk=1xjk-Zj(I)234117,:©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.(4).|Jc(I+1)-Jc(I)|,;,I=I+1,(2).,Jc.,Jc,.:,;;.,.,,.3C(CDMA),734140,99723,C13.9%;20052082005210(7d),().3.1(1),10(1).(2)CRISP2DM,SARS8.0K2Means,9().C1.,05,0,;5,;,,,.:2.5+2.5log(/)log(/):;.5,0,2.5.3.21,1,WAP.ARPU=47.3,1CTab.1ThecustomersegmentationofUNICOMbranchCnetwork1234567895.00.20.70.10.12.70.21.90.02.50.75.02.92.61.72.72.33.21.42.5IP1.92.65.02.01.52.62.12.71.62.50.94.73.95.02.13.42.93.41.02.59:0018:000.03.62.82.45.02.22.82.80.42.55.00.60.02.73.93.42.21.50.02.51.83.33.22.61.95.03.13.71.42.5WAP1.22.42.32.01.72.45.03.61.52.51.33.92.85.01.32.82.63.11.62.51.02.41.91.91.52.22.65.01.12.512324270825709050332943639273017813162799723/%12.42.72.69.133.43.62.71.831.7100ARPU/47.3511.9253.0209.474.4176.5134.5262.360.7104.2MOU/min292.2800.1774.8761.0243.9553.4354.8478.0155.0315.9ARPM/0.160.640.330.280.300.320.380.550.390.33/58.3138.665.0189.5247.664.236.746.7192.11038.7/%5.613.36.318.223.86.23.54.518.5100ARPU:;MOU:;ARPM(ARPU/MOU):;:;:441141©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.=511.9,4.91.,.3IP,..ARPU=253,2.43.IP,.4,.ARPU=209.4,2.01.,.5,9:0018:00,.ARPU=74.4,71%.,.6.ARPU=176.5,1.69.,.7WAP.ARPU=134.5,1.29.WAP,.8.ARPU=262.3,2.52.,.9,.ARPU=60.7,58%.,.4,,,.K2Means,.,,.,,,.,,K2Means,.,,,,.,,,,.,,,.:[1]WilliamL.Lifestyleconcepts[M].Chicago:Ameri2canMarketingAssn,1963.[2]WilliamW,DougT.Activities,interests,andopinion[J].JournalofAdvertisingResearch,1971,11(8):27-35.[3]HughesA.Strategicdatabasemarketing:Themasterplanforstartingandmanagingaprofitable,customerbasedmarketingprogram[M].NY:McGraw2HillProfessional,2000.[4]ClaudioM.Apracticalyetmeaningfulapproachtocustomersegmentation[J].JournalofConsumerMarketing,1981,15(5):494-504.[5]HaleyR.Benefitsegmentation:Adecision2orientedresearchtool[J].JournalofMarketing,1968,32(7):30-31.[6],.[J].,2006,20(1):12-14.LIUYing2zi,WUHao.Asummarizationofcustomersegmentation2methods[J].JournalofIndustrialEngi2neeringandEngineeringManagement,2006,20(1):53-57.54117,:©1994-2007ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.