Research On Credit Risk Assessment Of Enterprises Based On Support Vector Machine

Due to recent financial crisis and regulatory of BaselⅡ, credit risk assessment is becoming one of the most important topics in the field of financial risk management. The prevention of financial risk has been the primer problem in development of international financial markets. Commercial banks play an important role in the financial industry in China. Credit risk of commercial banks has become the most important and concentrative financial risk in China, which has great influence on the national economy. In the redit risk management process, the credit risk assessment of credit risk plays a vital link in the core of the special role.(1)This paper focuses on the loan enterprises which affect the credit risk of commercial banks. Through the study of financial factors of business loans, the goals to assess financial credit risk will be achieved. Support Vector Machines is used to complete the theoretical analysis and empirical studies in the future factors which affect the financial situation of business loans on the basis of the past research results. The main contents are as follows:(2)Taking loan enterprises which affect the credit risk of commercial banks as researching objection, analyze financial factors and non-financial factors of loan companies systematically.(3)The original data will inevitably lead to forecasting error in the results of empirical research without data pre-processing. This paper analyzes data factors through the significance test and factor analysis. Select input variables by significance test and compared the model with these input variables with the model using main factors as the input features.(4)Credit risk assessment study still stay in the traditional ratio analysis stage in our country, far from being able to meet the needs of credit risk decision. The evaluation model based on ensemble SVM model is constructed. And compared with the traditional MDA and Logistic methods, SVM model shows better results. This research results enrich and improve the credit assessment theory and methods. And has certain theoretical and practical significance to enhance credit risk prevention capacity, improve asset quality and profitability of credit.