Research On Credit Evaluation Of P2P Network Loan Borrowers Based On BP Neural Network

P2P(Online Peer to Peer loan),refers to a brand new loan mode,which use the Internet platform as a medium.On this platform the borrower and the lender make the funds trade happens,borrower needs to repay principal and interest within the prescribed time limit.This kind of loan mode subvert the traditional regional folk lending,break the fixed thought that lending must relies on relationships.P2 P loan starts in 2005,which mainly distributed in Europe and America.Several P2 P companies such as Zopa,Lending Club,Prosper and several P2 P are representatives,then set off a wave of net loan in the word.In 2007 P2 P net loan appears in China.This new mode of financing loans quickly without providing collateral,satisfying the needs of both funds side,making service areas which banks can't involve such as small and micro businesses financing and Individuals investment.P2 P net loan have a great impact on the traditional financial system in all aspects of the financial market that can not be ignored.However,for the recent years,the number of problem P2 P platforms did not decrease,there are still many platforms exposed running away,the borrower overdue is too serious,there is a shortage of funds and manager make serious dereliction of duty.P2 P net loan as a thing in the process of development,has a great uncertain risk in the future.As a basis for the healthy operation of P2 P net loan platform,the borrower's personal credit evaluation is a sub component of the entire loan project overall credit rating.Therefore,this paper mainly analyze the domestic and foreign evaluation methods,in base of domestic and foreign personal credit rating reference and different methods,combining with the characteristics of domestic P2 P net loan,select the borrower credit evaluation index and at last establishes the corresponding evaluation system.The work is to bulid up two kinds of BP neural network model according to the index system,and then compared the pros and cons of the model,select the better model for the credit evaluation of the borrower,which will take more accurate result of borrower's credit evaluation for P2 P net loan platform,so that the platform is smaller in the process of operation affected by the risk factors,to ensure the healthy development of P2 P net loan platform.The data selected from the recent four highest ranked platform P2 P net loan platform,the accuracy and applicability of the model verification results show that the BP neural network model using LM algorithm compared to the standard gradient descent algorithm model has better prediction ability,the output results are more accurate,but also can play a predictive role in the absence of information,the use of the model can be better used on the borrower's credit evaluation,so that the borrower's credit risks through platform operation can be obviously reduced.