基于logistic回归的违约概率模型的建立及分析
发布时间：2018-07-15 09:41
【摘要】：由于我国银行的商业化改革刚刚起步,法律、法规、数据和管理方法等还很不完善,这些都是产生金融风险的因素,其中信用风险是我国商业银行面临的最主要风险.我国加入WTO后,金融市场进一步对外开放,当前我国商业银行面临的主要任务是如何采取积极有效地措施与国际银行业接轨,这些任务首当其冲的是如何管理信用风险.《新巴塞尔协议》的实施要求银行建立完整的内部信用评级体系,对客户进行信用评级,量化贷款客户的信用风险.本文通过建立违约概率模型来预测客户违约情况,进一步加强商业银行信用风险的管理. 本文中的违约概率模型是基于Logistic回归分析。首先,采用逐步向后选择法进行模型指标的筛选。其次,通过Logistic回归得到违约概率模型.最后,我们利用Kolmogorov-Smirnov检验(简称K-S检验)与ROC曲线来检验模型区分违约客户的能力和准确度,最后确定违约概率模型.并且通过某商业银行的真实数据来对我们所建立的模型进行实证分析研究,得出结论.
[Abstract]:As the commercial reform of Chinese banks has just started, the laws, regulations, data and management methods are still very imperfect, these are the factors that produce financial risks, among which the credit risk is the most important risk faced by commercial banks in our country. After China's entry into WTO, the financial market is further opened to the outside world. At present, the main task facing our commercial banks is how to take active and effective measures to connect with the international banking industry. The implementation of Basel II requires banks to establish a complete internal credit rating system, credit rating for customers, and quantify the credit risk of loan customers. In this paper, the default probability model is established to predict the customer default and further strengthen the credit risk management of commercial banks. The probability model of default in this paper is based on Logistic regression analysis. First, the stepwise backward selection method is used to screen the model indexes. Secondly, the probability model of default is obtained by logistic regression. Finally, we use Kolmogorov-Smirnov test (K-S test for short) and ROC curve to test the model to distinguish the ability and accuracy of default customers, and finally determine the default probability model. And through the real data of a commercial bank to make an empirical analysis of our model, draw a conclusion.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.33
本文编号：2123650
[Abstract]:As the commercial reform of Chinese banks has just started, the laws, regulations, data and management methods are still very imperfect, these are the factors that produce financial risks, among which the credit risk is the most important risk faced by commercial banks in our country. After China's entry into WTO, the financial market is further opened to the outside world. At present, the main task facing our commercial banks is how to take active and effective measures to connect with the international banking industry. The implementation of Basel II requires banks to establish a complete internal credit rating system, credit rating for customers, and quantify the credit risk of loan customers. In this paper, the default probability model is established to predict the customer default and further strengthen the credit risk management of commercial banks. The probability model of default in this paper is based on Logistic regression analysis. First, the stepwise backward selection method is used to screen the model indexes. Secondly, the probability model of default is obtained by logistic regression. Finally, we use Kolmogorov-Smirnov test (K-S test for short) and ROC curve to test the model to distinguish the ability and accuracy of default customers, and finally determine the default probability model. And through the real data of a commercial bank to make an empirical analysis of our model, draw a conclusion.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.33
【参考文献】
相关期刊论文 前10条
1 宋冬梅;沈友娣;;运用Logistic模型评价上市公司信用风险[J];财会月刊;2008年05期
2 甄士龙;黎艳;;用Logistic模型估计企业的违约概率[J];广西金融研究;2008年11期
3 马永波;晏国祥;;内部评级违约概率度量模型的演进与启示[J];石家庄经济学院学报;2006年02期
4 蔡海燕;;商业银行信用风险管理模型应用探讨[J];经济师;2008年07期
5 林林;徐翔宇;;信用风险模型综述及对我国借鉴[J];金融经济;2009年02期
6 张燃;用Logit模型预测银行客户的违约概率[J];金融教学与研究;2005年04期
7 李越;;现代信用风险度量方法与模型述评[J];辽宁经济;2008年04期
8 赵中华;刘梅;;商业银行信用风险的VAR度量分析[J];商场现代化;2008年13期
9 柏艺益;;Logistic模型在评价上市公司信用风险中的应用研究[J];时代经贸(中旬刊);2008年S6期
10 管七海,冯宗宪;信用违约概率测度研究:文献综述与比较[J];世界经济;2004年11期
，本文编号：2123650
本文链接：https://www.wllwen.com/guanlilunwen/huobilw/2123650.html
