Comparative Research On Credit Risk Models Of P2P Network Lending

In recent years,with the increasing expansion of the Internet economy,has deeply affected various sectors of the society,P2 P network lending as an important model of the Internet economy has been rapidly developed.Compared with the traditional lending mode,P2 P network lending provides a new investment channel for lenders and provides a financing channel for borrowers to meet their needs for funds more conveniently and quickly.And the P2 P network lending is convenient.However,due to the virtual nature of the Internet it and the existence of asymmetric information between the two sides of the borrower,the borrower credit risk increases,and the lender cannot accurately judge the size of the risk of the borrower loan.Furthermore,it hinders the normal development of P2 P network lending market.In this paper,we choose the data of American P2 P network loan platform,and use a variety of models to predict the borrower credit risk in P2 P network lending.It provides some reference for the lender and P2 P network lending platform in how to reduce the risk of credit risk that the borrower may have.In this paper,we choose the data from 2017 loan records of Lending Club Corporation,a P2 P network lending platform in the United States,and use Logistic regression model,Probit regression model,Linear discriminant analysis and Random forest model based on integrated learning to study the borrower credit risk in P2 P network lending.Firstly,this paper introduces several models of P2 P network lending by combing the relevant literature,and expounds the angle and method of studying P2 P network lending.Then the basic idea of these four methods is expounded,then the original data set of loan collected is preprocessed,and then the pre-processed data set is prepared for the following modeling work.Four models are used to evaluate and predict the credit risk of the borrower in P2 P network lending.In the last step,using related evaluation index evaluate and summarize the prediction and analysis results of the four models.The results show that the Random forest model is better than other models in predicting and evaluating the borrower credit risk in P2 P network lending.The research also provide reference value to domestic P2 P network lending platform to the borrower credit risk assessment.It provides some suggestions for credit risk control mechanism in P2 P network lending,which help the lenders to guarantee legitimate income and promote the healthy and stable development of P2 P network lending market.