Research On Credit Risk Assessment Methods Of Borrower Of Commercial Banks In The Interest Rate Marketization

With the deepening of the Interest Rate Marketization, credit risk of borrower faced by commercial banks is becoming increasingly difficult to manage and control. Now, establishing accurate credit risk assessment model for commercial banks to assess borrower’s credit risk is particularly important.Under the premise of literature summarize about credit risk assessment model of domestic and foreign commercial bank, analyzed features of assessment model, ultimately choose support vector machine as a theoretical model of this article combined with China’s actual conditions. First, we use vector autoregression model and impulse response function to explore the dynamic relationship of the actual quarterly loan rate and the non-performing loans of joint-stock commercial bank between 2003 and 2013, which verifies a positive relationship between the actual loan rate and the non-performing loan ratio of the joint-stock commercial bank, so it’s necessary and urgent to establish credit risk assessment model in the interest rate marketization. The second, based on the financial indicators of listed companies as samples of borrowers, after select the training and testing samples, use principal component analysis to reduce dimensions of 20 indicators, then use support vector machine to establish credit risk evaluation model. This method prediction accuracy is as high as 94.4%, which is to assess the credit quality of commercial bank customers from a financial point of view.The credit risk assessment model established in this paper seeks to identify an effective method of credit evaluation, this classification method can provide an objective quantification of strong support for the credit department. It could help commercial banks as well as other financial institutions to construct credit risk assessment system of borrowers, to improve risk management capabilities of credit business, and to reduce the loss of non-performing assets. This work could also provide effective support for credit career and micro finance to develop better.