利率市场化下商业银行信贷客户信用风险评估方法研究

发布时间：2018-03-18 04:19

本文选题：利率市场化　切入点：商业银行　出处：《南京大学》2014年硕士论文　论文类型：学位论文

【摘要】：随着利率市场化的深入,商业银行面临的客户信用风险变得越来越难以管理和控制,此时,构建精准的信用风险评估模型来评估信贷客户的信用风险,对我国商业银行来说就显得尤为重要。本文在总结国内外关于商业银行信用风险评估模型文献的前提下,分析了已有评估模型的特点,结合我国实际,选择支持向量机作为本文的理论模型。首先,运用向量自回归模型和脉冲响应函数,探索我国2003年-2013年各季度实际贷款利率与商业银行不良贷款率之间的动态关系,验证了实际贷款利率与商业银行不良贷款率成正向关系,从而说明在利率市场化下商业银行建立信用风险评估模型是必要的和紧迫的；然后,以我国上市公司的财务指标作为银行信贷客户样本,选取训练样本和测试样本,先对20个指标采用主成分分析进行降维,再运用支持向量机模型进行信用风险评估,从银行客户财务角度识别银行客户信用质量,评估正确率达到94.4%,说明支持向量应用在信用评估领域是有效和精准的。本文建立的商业银行信用风险评估模型旨在寻求一种有效的信用评估识别方法,使其分类识别方式能为信贷业务部门提供客观有力的量化支持,帮助我国商业银行乃至其他金融机构很好的开展公司信用风险评估体系建设工作、提高信贷业务的风险管理水平、降低不良资产流失,为更好做强信贷事业及小微金融提供支持。
[Abstract]:With the deepening of interest rate marketization, the customer credit risk faced by commercial banks becomes more and more difficult to manage and control. At this time, the accurate credit risk assessment model is constructed to evaluate the credit risk of credit customers. It is very important for the commercial banks of our country. On the premise of summarizing the domestic and foreign literatures on the credit risk assessment models of commercial banks, this paper analyzes the characteristics of the existing models and combines the reality of our country. The support vector machine (SVM) is chosen as the theoretical model of this paper. Firstly, using the vector autoregressive model and the impulse response function, the dynamic relationship between the actual loan interest rate and the non-performing loan rate of commercial banks in each quarter from 2003 to 2013 is explored. The positive relationship between the actual loan interest rate and the non-performing loan rate of commercial banks is verified, which shows that it is necessary and urgent for commercial banks to establish a credit risk assessment model under the marketization of interest rates. Taking the financial index of Chinese listed companies as the sample of bank credit customer, the training sample and test sample are selected. The principal component analysis (PCA) is used to reduce the dimension of 20 indexes, and then the support vector machine (SVM) model is used to evaluate the credit risk. Identify the credit quality of bank customers from the point of view of bank customer finance, The accuracy of evaluation is 94. 4, which shows that the application of support vector is effective and accurate in the field of credit evaluation. So that the classification and identification method can provide objective and quantitative support for credit business departments, and help our commercial banks and even other financial institutions to carry out the construction of corporate credit risk assessment system. Improve the risk management level of credit business, reduce the loss of non-performing assets, and provide support for better credit and micro-finance.
【学位授予单位】：南京大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F832.4;F832.5

【相似文献】