Research On Credit Risk Assessment And Regulation Of China's P2P Online Loan Platform

Peers-to-Peers has entered a new stage of development under the situation of increasingly strengthened national supervision.As a new industry format in the development of Internet finance,P2 P is critical to the realization of inclusive finance in the era of big data.The history of large-scale credit risk explosion in the industry highlights the importance of credit risk identification and assessment in the booming trend of Internet finance.This article has combed the business logic of P2 P in detail.The development of P2 P has revealed the problems of the platform itself and the regulators.It is believed that the current business of more than 300 P2 P platforms in China has shown "small" and "decentralized" characteristics This is a good start for the management and control of credit risk.By analyzing the key factors of P2 P credit risk and the transmission mechanism,it is believed that the main prevention and control of credit risk should fall on the two largest sources of credit risk,namely the borrower and the platform itself,and must be supplemented by the scientific supervision of the regulator to be timely discover and mitigate credit risk.Based on this research,the scientific issues are as follows: First,cut from the borrowing side to reduce the business default risk level,and reduce the overall platform default risk by accurately identifying the target users.Second,cut the level of credit risk from the platform itself,and build a comprehensive evaluation model of credit risk for P2 P platforms.Third,put forward suggestions on effective intervention of regulatory management to reduce credit risk and build a scientific long-term P2 P online loan regulatory closed-loop system.For the prediction of user portraits,multiple algorithms such as machine learning and logistic regression are used to conduct research,and empirical prediction and analysis of individual users' gender,educational background,and age are conducted.Using 40000 pieces of user data on a platform,the results show that the decision tree algorithm has a better prediction of user education,and the algorithm of random forest combined with logistic regression has a higher accuracy of user gender prediction.In terms of constructing a comprehensive evaluation model of credit risk,this paper considers a gray linguistic variable method to build an evaluation system to evaluate platform credit risk from four aspects: economic and policy environment,platform customer credit status,platform credit management,and risk induction.By constructing gray linguistic variable algorithm,the platform credit risk assessment model is implemented and simulation trial calculation examples are given to verify the feasibility of the model,and five credit risk levels are divided correspondingly to the final calculated value.Regarding the P2 P online loan supervision recommendations,it is proposed that the macro level can be based on the gray linguistic variable credit risk assessment method proposed in this article,and the credit risk assessment of each platform can be dynamically and regularly performed,and the platform management measures of different risk levels can be designated.Under the guidance of the self-regulation convention,the Industry Self-Regulation Association improves the service functions of members and encourages industry innovation under the sandbox supervision.The platform can use its own data accumulation and data mining technology to realize the user portrait prediction algorithm and achieve the initial screening of high-quality customers.