基于人工神经网络的商业银行信用风险评估模型研究

发布时间：2019-05-29 02:38

【摘要】：经济全球化的趋势不断增强，国内推动利率市场化的步伐逐渐加快，金融市场的波动性日渐加剧，商业银行面临着前所未有的信用风险挑战。面对竞争日益激烈的生存环境，能否对贷款企业的信用风险进行科学、有效的管理对商业银行的可持续发展具有至关重要的影响。当前，我国的商业银行业尚处于改革转轨和新兴发展阶段，对信用风险的管理仍停留在传统的主观分析方法运用阶段，难以满足商业银行的发展需要。对此，本文站在商业银行的视角上，利用神经网络技术对商业银行贷款企业客户的信用风险评估问题进行研究，以期为商业银行提供一种有效的信用风险评估技术和方法。本文在对商业银行企业客户的信用风险评估模型文献研究的基础上，界定了商业银行信用及信用风险的内涵，运用人工神经网络基本理论，科学研究了商业银行信用风险评估的主要影响因素，构建了包含3个层次27个指标的商业银行信用风险评估指标体系，，通过综合比较研究商业银行信用风险评估模型，选择基于改进的BP神经网络建立了商业银行信用风险评估模型。利用搜集的144家公司的数据运用MATLAB2012a统计软件对该模型进行了实证分析。结果显示所构建的商业银行信用风险评估模型判别准确率高达87.04%，优于标准BP神经网络模型及Logistic回归模型的判别准确率，说明该模型对商业银行企业客户的信用风险能够进行合理、有效的评估。本文的研究成果对商业银行企业客户的信用风险评估提供了有益的评价依据，在实践工作中，具有较好的指导价值。
[Abstract]:With the increasing trend of economic globalization, the pace of promoting interest rate marketization in China is gradually accelerated, the volatility of financial markets is becoming more and more serious, and commercial banks are facing unprecedented credit risk challenges. In the face of increasingly fierce competition in the living environment, whether the credit risk of loan enterprises can be scientifically and effectively managed has a vital impact on the sustainable development of commercial banks. At present, the commercial banking industry of our country is still in the stage of reform, transition and new development, and the management of credit risk is still in the application stage of the traditional subjective analysis method, which is difficult to meet the development needs of commercial banks. In this paper, from the perspective of commercial banks, this paper uses neural network technology to study the credit risk assessment of commercial bank loan enterprise customers, in order to provide an effective credit risk assessment technology and method for commercial banks. Based on the literature research on the credit risk assessment model of commercial bank enterprise customers, this paper defines the connotation of commercial bank credit and credit risk, and uses the basic theory of artificial neural network. This paper scientifically studies the main influencing factors of credit risk assessment of commercial banks, constructs a credit risk assessment index system of commercial banks with 27 indicators at three levels, and studies the credit risk assessment model of commercial banks through comprehensive comparison. The credit risk assessment model of commercial banks is established based on the improved BP neural network. The model is empirically analyzed by using the data of 144 companies and MATLAB2012a statistical software. The results show that the discriminant accuracy of the credit risk assessment model of commercial banks is 87.04%, which is better than that of standard BP neural network model and Logistic regression model. It shows that the model can evaluate the credit risk of commercial bank customers reasonably and effectively. The research results of this paper provide a useful basis for the credit risk assessment of commercial bank customers, and have a good guiding value in practice.
【学位授予单位】：长沙理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.33;TP183

【参考文献】