个人通用信用评分领域中客户分组的国际经验与方法

International Experience and Methods of Customer Segmentation in the Field of Personal General Credit Scoring

客户分组(Segmentation)是个人通用信用评分建模过程中的重要环节,是指根据客群的行为差异将全部借款人划分为多个子客群,进而分别针对不同子客群进行信用评分模型的开发,相较于使用全样本构建的单一信用评分模型,良好的客户分组往往可以提升违约风险识别的准确率。美国三大个人征信机构益博睿(Experian)、艾克飞(Equifax)和环联(TransUnion)采用了费埃哲(FICO)和华帝评分(VantageScore)两种信用评分方法,通过跟踪费埃哲、华帝评分在个人信用评分领域中所使用的客户分组方法及其进展,为我国个人通用信用评分模型的开发提供经验借鉴。

Customer segmentation is an important step in the modeling process of personal general credit scoring,which refers to dividing all borrowers into multiple sub-customer groups according to the behavior differences of customer groups,and then developing credit scoring models for different sub-groups.Compared with the single credit scoring model constructed with full samples,good customer segmentation can often improve the accuracy of default risk identification.Experian,Equifax and TransUnion,the three major personal credit reporting institutions in the United States,have adopted two credit scoring methods:FICO and VantageScore.By tracking the customer segmentation methods used by FICO and VantageScore in the field of personal credit scoring and their progress,it provides experience for reference for the development of personal general credit scoring model in China.