Research On Credit Risk Rating Of P2P Network Loan Platform

As an important part of The Internet finance,P2P network lending develop very rapidly in China,and the number of platforms which offer services to peolpe who want to load by Internet increase a lot.However,under the condition of lack of supervision,the problem platforms is endless and has a huge negative impact on the investors and the development of whole industry.This paper introduces the credit risk rating method based on the credit risk rating theory and the credit risk rating method of the enterprises and the banks,the Credit Metrics model is chosen as the research method to establish the credit risk transfer matrix.Analyze the operation mode and profit model of the Internet loan platform,and combine the evaluation system of the rating agency to select the index of credit risk which reflects the loan platform.we use the clustering analysis method in unsupervised learning and the decision tree,SVM and lifting algorithm in supervised learning to measure the credit risk of net loan platform respectively.When do not know whether the loan platform is in breach of contract or not,we choise the cluster analysis method to get risk rating of platform,and track the results.The results show that about 7%of the all loan platforms are in breach of contract.The risk rating of those platforms are lower and the borrowing rate is negatively correlated with the rating results.After getting the problem platform sample,we use the clustering analysis,SVM and other algorithms to identify the credit risk of platforms.The results show that the supervised learning algorithm is superior to the clustering analysis.At the same time,these algorithms can predict the credit risk of the net loan platform,and the accuracy of the Adaboost algorithm is highest.Using the principal component analysis to classify the same net loan platform samples,the accuracy rate is lower comparing with the cluster analysis.Using Adaboost algorithm to study the credit risk of borrowers on net loan platform to supplement the credit risk rating of net loan platform.Finally,the credit risk transfer matrix of the net loan platform is established by using the credit risk rating result obtained by the lifting algorithm.At the same time,the default rate is added to characterize the repayment ability of the platform after default.Credit risk transfer matrix characterizes the default probability of loan loan platform.Combined with the recovery rate,we can estimate the credit risk of network loan platform size.The results show that the size of interest rates and the results of the rating was reversed,the lower the borrowing rate,the higher the credit rating of the loan platform.we can estimate the default risk of net loan platform and the entire industry by Credit rating on the network loan platform,which make regulators can better manage network loan the risk of network loan industry,and give investors some investment advices.