资源描述:
《小微贷授信额度测算的研究》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、上海交通大学硕士学位论文ESTIMATIONOFCREDITLIMITFORMICROCREDITABSTRACTMicrocreditisshort-term,small-scale,frequentandurgentinitscharacter.Currently,microcreditsaremostlycreditedbycreditmanagersandcreditcommitteecasebycaseinbanks,whichisslowandcostly,alsolackingofunifiedrules.Waytosol
2、vetheproblemistoreplacethecurrentbusinessstylebyintroducingmathematicsmethodsintotheestimationprocessofcreditlimitformicrocredit,makingtheprocessinformativeandstandardized.Byfittingthebehaviorpatternsofloanapplicantsintomathematicsmodels,itismorequick,preciseandlesscostl
3、yforbankstomakedecisionsonthenextcase.Twokindsofmodelswereusedtomakeastudyontheestimationofcreditlimitformicrocreditbasedonthephysicaldata,whichwereOrdinalLogisticRegressionmodelandFussyC-meansClusteringalgorithm.EmpiricalanalysisshowedthatOrdinalLogisticRegressionmodelw
4、asbetterthanFussyC-meansClusterinfittingwiththephysicaldatausedinthisarticle.Asadeductivemethod,theregressionmodelplaysgoodinimitatingtheexpertsinevaluatingtheapplicants’abilitytopaybacktheloan.Whilethefussyclustermodelisunsupervised,whichperformednotaswellastheregressio
5、nmodelwhenbasedonthedatacomingfromtheprophaseofthecreditprocess.Besides,OrdinalLogisticRegressionmodelisrigorousandforecastspreciselyandeasytobuild,whichmakesitagoodsolutionforcreditlimitestimation.Whilethemodelcameupinthearticlehelpsinimprovingitsefficiency,theestimatio
6、nofcreditlimitsformicrocreditsstillremainsaproblembecauseofthequalityofrawdatawasfarfromsatisfactory.Toomuchnoiseinthedataconstrainstheactualresultofthemodel.II万方数据上海交通大学硕士学位论文Standardizationindatacollectionandfindingwaystoreducenoisemeansalot,whichwasalsodiscussedinthel
7、astpartofthearticle.KEYWORDS:Microcredit,OrdinalLogisticRegressionmodel,FussyC-meansClusteringmodelIII万方数据上海交通大学硕士学位论文目录第一章绪论···················································································································11.1选题背景和意义····················
8、··············································································11.2研究思路·················