Identification And Measurement Of Credit Risk Of Peer-to-student Lending Based On Customer Segmentation

In recent years,P2P small loan model has developed rapidly.As of the end of November 2016,China has normal operation P2P network loan platform as high as 2534.The industry competition has intensified,and the industry market began to subdivide.In many of the market segments,the college students P2P network loan develops rapidly.Compared with other net borrowers,college students have many advantages,such as identity authentication?high cost of breaking contracts?parents'potential guarantee.The credit risk is relatively smaller.However,there are still a lot of overdue defaults on college students' net loan platform.This is mainly due to net loan are so simple and convenient that leads to impulsive consumption.As well as college students do not have a fixed source of income,and credit awareness is not strong.This paper is mainly to plan and implement a credit loan identification and measurement program for college students' loan project.The implementation and verification of this program are based on the "instant 360" platform project information.The implementation of this program mainly includes four steps:First,based on the customer segmentation theory,the college students network loan items are divided into three categories:low education population,the general population and highly educated population.Secondly,for each group project,we use the XGBoost algorithm to identify the credit risk of each loan group.The result shows that the algorithm is suitable to credit risk identification,and the accuracy and stability of the subsequent methods is improved by using the feature selection method to reduce the dimension processing.Finally,the results of XGBoost algorithm and the scorecard model are used to measure and test the credit risk of the loan project.The implementation of this program is very good.It's a good credit risk identification program.