A Study On The Credit Risk Assessment Of Borrowers On P2P Lending Platforms

At present,the risks of P2P online loan industry are constantly expanding.Due to the borrower's default behavior,"out of control" and "thunder" problem platforms increase sharply,many P2P online loan platforms have serious solvency crisis,which hinders the sound development of P2P online loan platforms.Therefore,the identification and effective assessment of the borrower's credit risk level will be conducive to maintaining the interests of P2P online lending platforms and lenders,thus promoting the overall sound development of the online lending industry.At present,researches on risks of P2P lending industry in China are not in-depth enough,and most of them focus on qualitative research.The research perspective is mainly focused on the analysis of supervision level,which is relatively single Therefore,for the healthy development of P2P online loan industry,it is necessary to supplement the research from the perspective of borrower credit risk assessmentThis paper first introduces the status quo,operation mode of P2P lending platform and the theories related to the credit risk of borrowers on P2P lending platform,then summarizes the individual credit risk assessment model and the selection of individual credit risk assessment indicators.Then use python software write the crawlers,obtained”borrowed" everyone of the borrower's data,and the data preprocessing and descriptive statistical analysis,and finally to the borrower data using Logistic regression model and neural network model and screening algorithm based on neural network is an important variable of the Logistic regression model has carried on the empirical analysis,and using ROC curve and AUC values on the performance of the model testIn under the background of this paper,based on neural network algorithm screening of important variables of the Logistic regression model to improve the level of the borrower's credit risk assessment also retains strong explanation,the Logistic regression model that based on neural network algorithm screening of important variables of the Logistic regression model for P2P platform of borrower credit risk assessment in the field of better adaptability This paper suggests that P2P platforms should establish more effective ways of credit risk assessment,strengthen the construction of credit data sharing platforms,and incorporate them into the credit investigation system of the central bank as soon as possible to promote the healthy development of P2P industry.