基于数据挖掘的中小微企业信贷评估决策研究  

Research on Credit Evaluation Decision of Micro Small and Medium Sized Enterprises Based on Data Mining

作　　者：孙静懿[1] 冯由玲 罗庆 SUN Jing-yi;FENG You-ling;LUO Qing(Department of Cultural Foundation,Changchun Finance College,Changchun 130028,China;School of Management Science and Information Engineering,Jilin University of Finance and Economics,Changchun 130117,China)

机构地区：[1]长春金融高等专科学校文化基础教研部,吉林长春130028 [2]吉林财经大学管理科学与信息工程学院,吉林长春130117

出　　处：《长春金融高等专科学校学报》2021年第3期5-15,共11页Journal of Changchun Finance College

基　　金：吉林省教育厅项目(JJKH20210142SK)。

摘　　要：以中小微企业为研究对象,基于有信誉记录的中小微企业数据,通过K-means算法对企业的经济实力做聚类分析,构建风险系数衡量指标并进行回归分析,构建有信誉记录的信贷决策模型。基于无信誉记录的中小微企业数据,根据企业票据交易关系用聚类方法为企业分类,通过回归方程计算负票占有率、金额变动值和重复成交率之间的关系,对企业借贷能力进行评分,构建无信誉借贷决策模型,并在此模型基础上,通过回归方程对突发因素情况下企业有效利润进行计算,从而对企业借贷能力进行评分,构建突发因素无信誉借贷决策模型,为各类信贷机构的贷款策略提供参考。Taking micro small and medium-sized enterprises as the research object,based on the data of small and medium-sized micro enterprises with credit records,this paper uses k-means algorithm to do cluster analysis on the economic strength of enterprises,constructs risk coefficient measurement index and regression analysis,and constructs credit decision-making model with credit records.Based on the data of micro small and medium enterprises with no credit record,according to the transaction relationship of enterprise bills,this paper uses the clustering method to classify enterprises,and calculates the relationship among negative ticket occupancy rate,amount change value and repeated transaction rate through regression equation,so as to score the enterprise's lending ability,on the basis of this model,the effective profit of enterprises under sudden factors is calculated by regression equation,so as to score the enterprise's lending ability,and build the decision model of sudden factors'creditless lending,so as to provide reference for the lending strategies of various credit institutions.

关 键 词：决策模型 K-means聚类分析 回归模型 决策树 

分 类 号：F832.4[经济管理—金融学] F276.3