商业银行信贷客户违约预警建模分析

Warning System Models Analysis for National Commercial Banks against Defaults of Credit Clients

10.6338/JDA.200604_1(2).0001

白羽(Yu Bai)；宋逢明(Feng-Ming Song)；王凯(Kai Wang)

信贷风险 ； 多元判别分析 ； logistic回归 ； 决策树 ； 神经元网络 ； Credit Risk ； Multivariate Discrimination Analysis ； Logistic Regression ； Decision Tree ； Neural Network ； ROC

Journal of Data Analysis

1卷2期（2006 / 04 / 01）

1 - 13

簡體中文

本文应用多元判别分析、logistic回归、决策树和神经元网络等4种有代表性的分类方法，对国内两组不同时期、不同规模城市商业银行的实际数据集进行了全面的实证分析。在模型分析中，文章加入了虚拟变量来描述宏观经济和行业因素对违约的影响，并引入了收益总额矩阵度量模型绩效。结果表明，多元判别分析的判别能力最差，logistic回归方法也不太理想；而决策树方法和神经元网络方法比较理想，特别是决策树方法，从多个角度上分析所得到的结果都是最好的。因此，在国内商业银行对实际信贷客户进行违约预警建模型时，本文推荐使用决策树方法和神经元网络方法。

Based on the real data sets involving different periods and different sizes of cities from national commercial banks, this paper applies the methodologies of Multivariate Discrimination Analysis, Logistic Regression, Decision Tree and Neural Network to accomplishing the detailed empirical research. In the sample analysis, two dummy variables are adopted to describe the factors of macroeconomics and industries which will likely affect the states of defaults. Furthermore, this paper introduces ”income array” to measure the performance of the models. The results shows that the determinant power of Multivariate Discrimination Analysis is the most ineffective, and Logistic Regression is also unsatisfactory, while Decision Tree and Neural Network are better, especially the former, whose results are almost perfect in various analyses. Therefore, this paper recommends the methodologies of Decision Tree and Neural Network for national commercial banks to create models of a national warning system against defaults of credit clients.

1. Altman, Edward I.,Caouette, J.,Narayanan, P.(1998).Managing credit risk: The ironic challenge in the next decade.John Wiley and Sons, Inc.
2. Breiman, L.(1984).Classification and Regression trees.Monterey, CA:Wadsworth and Brooks.
3. Galindo, J.,Tamayo, P.(2000).Credit risk assessment using statistical and machine learning: basic methodology and risk modeling applications.Computational Economics,15(1-2)
4. Hsu, C.-W.,Lin, C.-J.(2002).A comparison of methods for multi-class support vector machines.IEEE Transactions on Neural Networks,13(2),415-425.
5. Huang, Z.(2004).Credit rating analysis with support vector machines and neural networks: a market comparative study.Decision Support Systems,37(4),543-558.
6. Johnson, R. A.,Wichern, D. W.(2002).Applied multivariate statistical analysis.Upper Saddle River, N.J.:Prentice Hall.
7. Johnson, Richard A.,Wichern, Dean W.(2003).Applied Multivariate Statistical Analysis.
8. Malliavin, P.,Mancino, M.(2000).Fourier Series Method for Measurement of Multivariate Volatilities.Finance and Stochastic,6,49-61.
9. Michel, C.,Dan, G.,Robert, M.(2000).A Comparative Analysis of Current Credit Risk Models.Journal of Banking and Finance,24,59-117.
10. Pinches, g.,Mingo, K.(1982).A multivariate analysis of industrial bond ratings.The journal of finance,1982(28),1-27.
11. Quinlan, J. R.(1993).C4.5 programs for machine learning.Morgan Kaufmann.
12. Ross, Stephen A.,Westerfield, Randolph W.,Jaffe, Jeffery F.(2002).Corporate Finance.McGraw-Hill Irwin.
13. Saunders. A.(1999).Credit Risk Measurement: New Approaches to Value at Risk and other Paradigms.John Wiley & Sons. Inc..
14. Swets, J. A.,Pickett, R. M.(1982).Evaluation of Diagnostic Systems: Methods from Signal Detection Theory.New York, NY:Academic Press.
15. 石月华(1993)。银行信用评估模式之建立。交通大?管理科?研究所。
16. 朱世武(2003)。SAS编程技术与金融数据处理。清?大?出版社。
17. 宋逢明(1999)。金融工程原理—无套利均衡分析。清?大?出版社。
18. 范金城、梅长林(2002)。数据分析。科?出版社。
19. 郭敏华(2004)。信用评级。中?人民大?出版社。
20. 郑杰(2002)。贷款风险分类的管理与应用。中?金融出版社。