Research On Loan Scoring Calculation Of P2P Lending Based On Logistic Regression

P2P lending is a product of traditional financial and Internet integration.Unlike traditional finance,P2P lending is a decentralized model that enables individuals to lend to individuals,ie,peer-to-peer,without the need for intermediaries such as financial institutions to participate.Borrowing in all aspects.Because the P2P platform has the characteristics of convenience and speed,compared with the financial institutions such as banks,the P2P platform has lower requirements for borrowers,and the lending process is more convenient.More and more people choose to borrow through the P2P platform.However,due to the characteristics of P2P network lending itself,it is especially important to control the risk of P2P platform,continuously reduce the default rate of borrowing,increase the rate of borrowing,and find the influencing factors of these two become the first platform of each P2P network lending platform.Objectives,therefore,studying the risk control mechanism of P2P platform has important practical significance.In the existing research,most of the scholars have studied the factors affecting the default rate and the rate of return,but they have combined the two to study.In view of this,based on the factors influencing the default rate,this paper further introduces the concept of loan yield,trying to combine the advantages of the two to divide the loan,and obtain a more effective loan screening mechanism to meet different Lenders with risk appetite,enriching research in the field of P2P network lending.This paper selects the data of 890,000 loan transactions between 2007 and 2015 on the Lending Club platform.After cleaning the data set,it retains 240,000 credit loan data as a sample of this study.The data set mainly includes information such as the borrower's personal information,loan information,borrower's credit information,loan status and repayment information.After descriptive statistics on the samples,multiple machine learning models were used to predict the default rate,and finally a logistic regression model with higher accuracy was selected.Then,the results of logistic regression are analyzed.The factors affecting the default rate are discussed.Another data mining technique xgboost is used to sort the features and compare them with the logistic regression results.Finally,the predicted default rate and the existing rate of return are combined to arrive at the loan score,which is further discussed.The results of data analysis show that:First,in the case of considering only the loan default rate,the borrower's annual income,total credit balance,debt-to-income ratio,platform credit rating and loan purpose have a greater impact on whether or not the default is breached.Second,at the same time Considering the default rate and the rate of return,the factors affecting the loan score and the factors affecting the default rate are similar,but the performance is different under different loan purposes.The default rate of the loan for the wedding is the lowest,and the average yield is the highest.However,the average loan benefit of the wedding,that is,the rating is not the best,and the highest is the car loan,this method provides a new perspective to judge the quality of the loan;third,it is not difficult to find out from the study,The online lending market is not a fully effective market,and there is still much room for improvement.This study broadens the research on the credit risk field of P2P platform.Specifically,firstly,according to the regression results,several factors that have a great influence on the default rate are obtained and explained,and the previous literature is supplemented.Second,adopting a new perspective,considering both the default rate and the rate of return,and numerically analyzing the influencing factors of the loan score,focusing on how the difference in loans under different loan purposes is a lending platform and lending individuals.Provides a valuable reference indicator for evaluation.The research conclusions of this study have important guiding significance for the P2P platform and lenders in selecting loans.According to the loan score and risk preference,the borrower is selected to lend,and the maximum profit can be realized while avoiding risks.Continuously eliminate inferior loans and increase the efficiency of the P2P lending market to develop into a fully efficient market.