Research On P2P Lending Credit Risk Assessment Based On Disequilibrium Fuzzy Proximal Support Vector Machine Model

P2P(Peer-to-Peer)Lending is a new private lending based on network platform.This model can achieve lending money derectly to others through Internet,so it has benefit to reduce the investment and financing cost,improve the return of investment and operate conveniently.However,due to the asymmetric information in P2 P lending,the investors can't get the full information of borrowers.Compare with traditional commercial bank,the loans in P2 P Lending doesn't have mortgage.And its risk management ability is far below traditional financial market.It makes the investors face a higher credit risk.Faced with this reality,this article considered the development of P2 P Lengding and establishs the specific indexes and model to evaluate the borrowers' credit risk in P2 P Lending.This paper studies from the following aspects:First of all,this paper introduced the development and operation model of P2 P Lending,and analyzed the credit risk of P2 P Lending.According to the informations which borrowers published in P2 P Lending platform and the influence on borrowers' credit risk,this paper established the specific indexes for the borrowers' credit risk research in P2 P Lending,and also picked the suitable indexes for P2 P Lending platform credit risk assessment.Next,this paper introduced the fundamental and algorithm features of SVM,PSVM and FPSVM,and analyzed the applicability of the SVM in P2 P Lending personal credit risk assessment.Considering the disequilibrium distribution on trusted users and default users in P2 P lending,and different acceptable degree of investors in classification error,this article introduce the bilateral weighted error measuring method based on fuzzy proximal support vector machine(FPSVM).In order to assessment the credit risk of P2 P Lending,we proposed the disequilibrium fuzzy proximal support vector machine(DFPSVM),and measure fuzzy memberships of the positive and negative samples error term through mapping distance.And then,we get the borrowers' credit scoring and credit rating model.Meanwhile,we established the P2 P Lending platform credit risk assessment model based on PCA-AHP.Then,this paper made empirical researches on evaluate the borrowers' credit risk in P2 P Lending based on disequilibrium fuzzy proximal support vector machine(DFPSVM)model.The borrowers' informations in Renrendai which obtained from web crawler technology are used to test and compare the performance of different models.The experimental results revealed that the proposed DFPSVM model has better generalization ability and higher classification accuracy than other existing models,such as LR,BPNN,SVM,PSVM and FPSVM.It can effectively reduce the effect of disequilibrium samples and provide decision support for investors in P2 P Lending.Futhermore,we used the model calculated the borrowers' credit score,and established credit rating of borrowers.It provides a more transparent borrowers' credit risk assessment method for investors,and help investors choose the borrowers which has higher credit rating.It can also provide a new credit rating method for P2 P Lending,and reduce the platform's credit risk.At last,to analyze the credit risk of P2 P Lending platform,we proposed PCA-AHP model to calculate the credit score of each P2 P Lending platform.In this section,we choosed ten famous P2 P Lending platform in China,and obtained the credit risk rank of these platform by used PCA-AHP model.Compared with the 2016 ranking of P2 P Lending platform in China published by Online Lending House,the PCA-AHP model can assessment the credit risk of P2 P Lending scientifically and reasonably.