Application Of Support Vector Machine In Credit Risk Assessment Of P2P Borrowers

In the financial industry,risk is an eternal theme.How to correctly assess the risks in the financial industry is of great significance to the development of finance.Since the emergence of the first P2 P platform in China in 2007,P2 P lending has begun to develop rapidly.However,at present,China's P2 P lending is still in the primary development stage.The supervision and system construction are not perfect,which has led to the emergence of many problem platforms and hindered the development of the P2 P lending industry.Compared with the traditional lending,the P2 P lending industry has not yet established a comprehensive credit risk assessment system and the risk assessment for borrowers is still immature,which leads to serious potential credit risk of borrowers and thus affects the development of the P2 P lending industry.It is important to conduct risk assessment on P2 P lending for making P2 P lending better.Support Vector Machine(SVM)is a machine learning algorithm based on statistical learning theory proposed in the 1990 s.It solves the optimization problem by finding the Optimal Separating Hyperplane and is applicable to solving nonlinear and small sample problems.Firstly,this paper expounds on the P2 P lending.Secondly,introducing the relevant theoretical knowledge of the support vector machine.Thirding,analyzing the causes of the credit risk of the borrower of the P2 P lending.Next,constructing an index system to evaluate the credit risk of the borrower.Finally,according to analyzing the characteristics of the research problem,a support vector machine model was established.The model was trained by using the processed data that is the borrower personal information obtained from the the peer-to-peer lending website.Based on the experimental results,it was found that:(1)Compared with the existing qualitative analysis of credit risk,it is feasible to turn the research on credit risk of P2 P borrowers into default classification for borrowers.(2)The performance of default classification with support vector machines is different while choosing different kernel functions.On the performance of borrower default classification,radial basis function has the best classification effect.(3)Different parameter choices will affect the accuracy of support vector machine in default classification.(4)Compared with the other algorithms,the default classification of support vector machine is superior to other algorithms.This paper uses a support vector machine for credit risk analysis.It transforms credit risk analysis into a classification problem and obtains a high classification accuracy rate.It can be concluded that not only traditional credit risk analysis methods can be used in the process of assessing credit risk.Machine learning algorithms can also be used to make risk assessment more comprehensive.