The Evaluation Of P2P Borrower's Credit Risk

In the era of information technology,"Internet plus" mode has led the entire industrial structure,the Internet finance is progressing quickly.As an important measure of economic restructuring,the arrival of P2P online lending is undoubtedly an effective tool.P2P online lending,based on the network platform,can accomplish financial exchanges and lending transactions between person and person.It gathers the small amount of funds to obtain greater benefits.As an emerging industry,although the country promulgate relevant laws and regulations constantly,there are still many problems to solve,the most important is the credit risk assessment of the borrowers.The research on this problem is of great importance to the P2P online lending platform,as well as for the investors which offers a very important practical value and guiding significance.This master's thesis systematically discusses the main operating modes,characteristics and related risks of domestic P2P online lending.As for the core of the P2P borrower's credit risk,according to the data collected by the P2P online lending platform use both qualitative and quantitative methods to scientifically select the main factors which affect the borrower's credit risk and determine the appropriate variables then construct the P2P borrower's credit risk assessment system eventually,at the same time use the method of descriptive statistic to analysis.What's more,with the quotation of two kinds of classification algorithms,decision tree and RBF neural network,establish P2P borrower's credit risk evaluation model by simulation training,to assess the borrower's credit risk effectively.Finally,compare the index selection,prediction accuracy with the outputs of the two algorithms and draw a conclusion.The four variables,interest rate,repayment period,total loan amount and the number of repayments are the common evaluation indexes of the two algorithms.The prediction accuracy of decision tree model with less evaluation index is higher than that of RBF neural network model considering all evaluation indexes.However,the RBF neural network model has the advantage of identifying the borrowers with good credit alone.