基于SVM的我国商业银行信用风险管理模型研究
发布时间：2018-10-11 09:57
【摘要】：信用在市场经济条件下具有非常重要的作用,信用是市场经济运行的前提和基础,它帮助资金和其他生产要素在经济体系内部流动,是整个经济的润滑剂。商业银行作为我国金融体系的重要组成部分,面临着各种风险,而信用风险是我国商业银行中最主要的风险之一。在经济全球化的背景下,行业内的竞争日益激烈,因此提高我国商业银行的信用风险管理能力至关重要。但是,由于信用风险不确定性及违约数据难获得的特点,我国长期以来对信用风险的分析停留在传统的历史财务比率分析和信用分析上,因此,找到一个准确度量、控制并管理信用风险是当今金融业的一个重点和挑战。 本文首先介绍了信用风险和信用风险管理的概念、研究背景和发展历程,然后介绍了目前信用风险管理的几种方法,并对其优缺点进行简单分析。第二部分重点介绍了支持向量机方法,介绍了理论基础和SVM方法的应用。第三部分,本文对模型进行了一些改进,从模型样本数据变量的选择、最优参数寻优等方面进行了改进,提高了模型的预测正确率。最后,通过某商业银行企业客户数据的测试表明,改进的支持向量机方法对于信用风险违约情况的预测正确率要高于传统的SVM算法。
[Abstract]:Credit plays a very important role in the market economy. Credit is the premise and foundation of the market economy. It helps the capital and other factors of production to flow through the economic system and is the lubricant of the whole economy. As an important part of our country's financial system, commercial banks are faced with various risks, and credit risk is one of the most important risks in our country's commercial banks. Under the background of economic globalization, the competition in the industry is becoming increasingly fierce, so it is very important to improve the credit risk management ability of commercial banks in China. However, because of the uncertainty of credit risk and the difficulty of obtaining default data, the analysis of credit risk in our country for a long time has stayed in the traditional historical financial ratio analysis and credit analysis. Controlling and managing credit risk is a key point and challenge of financial industry today. This paper first introduces the concept of credit risk and credit risk management, research background and development process, then introduces several methods of credit risk management, and analyzes their advantages and disadvantages. In the second part, the support vector machine (SVM) method is introduced, the theoretical basis and the application of SVM method are introduced. In the third part, some improvements are made to the model, such as the selection of sample data variables, the optimization of optimal parameters, and so on, so as to improve the prediction accuracy of the model. Finally, a commercial bank enterprise customer data test shows that the improved support vector machine method for credit risk default prediction accuracy is higher than the traditional SVM algorithm.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.33;TP18
本文编号：2263776
[Abstract]:Credit plays a very important role in the market economy. Credit is the premise and foundation of the market economy. It helps the capital and other factors of production to flow through the economic system and is the lubricant of the whole economy. As an important part of our country's financial system, commercial banks are faced with various risks, and credit risk is one of the most important risks in our country's commercial banks. Under the background of economic globalization, the competition in the industry is becoming increasingly fierce, so it is very important to improve the credit risk management ability of commercial banks in China. However, because of the uncertainty of credit risk and the difficulty of obtaining default data, the analysis of credit risk in our country for a long time has stayed in the traditional historical financial ratio analysis and credit analysis. Controlling and managing credit risk is a key point and challenge of financial industry today. This paper first introduces the concept of credit risk and credit risk management, research background and development process, then introduces several methods of credit risk management, and analyzes their advantages and disadvantages. In the second part, the support vector machine (SVM) method is introduced, the theoretical basis and the application of SVM method are introduced. In the third part, some improvements are made to the model, such as the selection of sample data variables, the optimization of optimal parameters, and so on, so as to improve the prediction accuracy of the model. Finally, a commercial bank enterprise customer data test shows that the improved support vector machine method for credit risk default prediction accuracy is higher than the traditional SVM algorithm.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.33;TP18
【参考文献】
相关期刊论文 前4条
1 王睿;;关于支持向量机参数选择方法分析[J];重庆师范大学学报(自然科学版);2007年02期
2 王兴玲,李占斌;基于网格搜索的支持向量机核函数参数的确定[J];中国海洋大学学报(自然科学版);2005年05期
3 黄青;彭家瑚;;我国商业银行信贷风险管理系统的构建[J];企业经济;2006年11期
4 李盼池,许少华;支持向量机在模式识别中的核函数特性分析[J];计算机工程与设计;2005年02期
相关硕士学位论文 前2条
1 郭成报;支持向量机最优参数选取及应用[D];山东大学;2011年
2 王艾婷;基于SVM的商业银行信用风险模型研究[D];天津大学;2009年
，本文编号：2263776
本文链接：https://www.wllwen.com/guanlilunwen/bankxd/2263776.html
