Research On Credit Risk Prediction Of Consumer Finance Based On Improved BP Neural Network

Domestic consumer finance has flourished in recent years,internet consumer finance has risen rapidly,such as private P2 P loan has greatly increased all over China.Since the end of 2016,the personal short-term loans has accelerated,with the proportion of consumer loans rising to80% and the loan size approaching 8 trillion.Because of a large number of consumer finance companies start P2 P loan business in recent years,which do not have strict risk control system,the traditional credit control from credit reference center the People's Bank of China to obtain credit information,which increase the cost.And also the traditional credit control can not grasp the borrower's overall debt situation,such as the borrower's debt on other platforms.On the other hand,due to the loan business mostly for personal customers,such as migrant workers,college students,self-employed people who has not yet registration in credit reference system,so that the traditional indicators for this kind of people are "ineffective",so that brings negative impact to the business development of consumer finance companies.The paper will studies the credit data of licensed consumer finance companies on behalf of H companies,using quantitative analysis and qualitative analysis to solve the problems and bottlenecks of traditional risk control models base on the big data.The complex risk control model problem is transformed into a data problem to reduce the uncertainty risk in the risk control model.In the case study,the influence factors and weights of credit risk control are mainly studied from the perspective of borrowers,and the effect and improvement measures of big data risk control model are studied from the perspective of lending institutions.The result shows that the relative error between the prediction result and the real credit situation is less than 2%,and the range of change is stable.Through the comparative analysis of the relative error,the accuracy and stability of the credit situation predicted by the index of e-commerce,social network and search behavior are proved.Finally,from the financial market regulators,consumer financial companies,individual customers three aspects of consumer credit risk control put forward suggestions and st rategies,intended to regulate customer credit behavior.