Applicability Research On Credit Risk Assessment Model Of P2P Online Loans

Online credit risk assessment is an important and realistic issue.It includes credit risk assessment of online lending platform and personal credit risk assessment of online lending.It is of theoretical and practical significance to evaluate the credit risk of online loans.This paper uses the theories of finance and econometrics,uses the credit risk assessment indicators and data of the P2 P online lending platform,and builds the credit risk assessment model of the online lending platform with reference to the Logit model,and conducts empirical analysis;combined with the big data of Lujin and Randomly extracted samples were established to establish Probit personal credit assessment model,linear regression and logistic regression combination model,and empirical research.The conclusion of the thesis is that the dependent variable index of the online loan platform risk assessment model is the platform level.The independent variable indicators are average interest rate,per capita borrowing amount,claims against the index,operating index,monthly active borrower,platform comprehensive score,before The credit risk of the 50 platforms is low,but the credit risk of the other platforms is higher;the dependent variable indicator of the personal credit evaluation model is whether to pay repayment,the independent variable is the age of the borrower ag,the proof of income ic,the certificate of real estate hc,Credit certificate pc,risk margin ip,loan amount is,overdue number os,the personal credit risk of Lu Jin Suo in the past two years is generally low;the online loan credit risk evaluation model should be optimized with the development of society,economy,culture and technology.The credit evaluation of online loan platform should adopt Logit model;The personal credit risk evaluation should adopt the Probit regression model when there are many subtypes of independent variables.When there are more continuous independent variables,the combination model of linear regression and logistic regression should be adopted.