基于数据挖掘的贷款信用风险评估方法比较研究
|
摘要
东南亚金融危机的爆发使得全球金融市场之间紧密联系、不同市场之间的互相影响、银行风险和金融危机在国际间的传播问题开始引起越来越多的关注。目前,许多金融机构陷入经营困境的主要原因不再是信用风险或市场风险等单一风险,而是由信用风险、市场风险外加操作风险相互交织、共同作用造成的。
巴塞尔委员会于1998年推出《有效银行监管的核心原则》,至此,市场风险与信用风险、操作风险一并成为银行监管部门重点关注的对象。2005年11月,巴塞尔委员推出《新巴塞尔资本协议》,新资本协议对银行信用风险管理提出了更高的要求,对数据和量化方法提出了更高的要求。虽然我国商业银行对所面临的市场风险的认识程度有了一定的提高,同时监管部门的监管水平也有了较大程度的提高。但对商业银行所面临的信用、操作和市场这三大主要风险的研究和实践在我国尚处于摸索阶段。个人信用是市场经济的基础;没有个人信用,市场经济便失去了坚实基础。
本文以银行贷款的信用风险为核心,主要研究了不同的数据挖掘方法在商业银行贷款信用风险评估中的性能,包括神经网络算法、遗传算法和决策树算法。首先,我们分别用这三种算法构建信用风险评估模型,然后通过模型对贷款数据进行分析,得出信用风险评估结果。最后,对三种模型的测试结果进行比较分析,研究不同的算法在信用风险评估中的性能,得出本文的结论,为银行贷款人员进行贷款信用风险分析提供决策依据。
Since the 1997 Southeast Asian financial crisis, people have paid more attention to the close contact of global financial markets, mutual influence among different market, banking risks and financial crisis. At present, the main reason why many financial institutions get bogged down into operating difficulties is no longer caused by single risk such as market risk, credit risk, but by the plus intertwined, common function of credit risk, market risk and operational risk.
Basel Committee launched "the core principles of effective banking supervision" in 1998, at this point, market risk and credit risk, operational risk become the object on which banking supervision departments focus. November 2005, the Basel Committee launched "New Basel Capital Accord". New Capital Accord puts higher requirements on bank credit risk management and data and quantitative approach. Although Chinese commercial banks have raised the awareness about market risks and the supervision levels supervision departments have also improved, the research and practice on credit, operational and market risk is still in the exploratory stage. Personal credit is the basis of market economy; without personal credit, the market economy will lose solid foundation.
This paper, with bank loan credit risk as the core, mainly studies the performance of different methods of data mining in commercial bank loan credit risk assessment, including neural network algorithm, genetic algorithms and decision tree algorithm. First of all, we use the three algorithms to construct credit risk assessment model separately, and then analyze the loan data with the models we have constructed and get the results of credit risk assessment. Finally, we compare and analyze the test results of the three models, then study the strengths and weaknesses of different algorithm in credit risk assessment and come to the conclusion.
巴塞尔委员会于1998年推出《有效银行监管的核心原则》,至此,市场风险与信用风险、操作风险一并成为银行监管部门重点关注的对象。2005年11月,巴塞尔委员推出《新巴塞尔资本协议》,新资本协议对银行信用风险管理提出了更高的要求,对数据和量化方法提出了更高的要求。虽然我国商业银行对所面临的市场风险的认识程度有了一定的提高,同时监管部门的监管水平也有了较大程度的提高。但对商业银行所面临的信用、操作和市场这三大主要风险的研究和实践在我国尚处于摸索阶段。个人信用是市场经济的基础;没有个人信用,市场经济便失去了坚实基础。
本文以银行贷款的信用风险为核心,主要研究了不同的数据挖掘方法在商业银行贷款信用风险评估中的性能,包括神经网络算法、遗传算法和决策树算法。首先,我们分别用这三种算法构建信用风险评估模型,然后通过模型对贷款数据进行分析,得出信用风险评估结果。最后,对三种模型的测试结果进行比较分析,研究不同的算法在信用风险评估中的性能,得出本文的结论,为银行贷款人员进行贷款信用风险分析提供决策依据。
Since the 1997 Southeast Asian financial crisis, people have paid more attention to the close contact of global financial markets, mutual influence among different market, banking risks and financial crisis. At present, the main reason why many financial institutions get bogged down into operating difficulties is no longer caused by single risk such as market risk, credit risk, but by the plus intertwined, common function of credit risk, market risk and operational risk.
Basel Committee launched "the core principles of effective banking supervision" in 1998, at this point, market risk and credit risk, operational risk become the object on which banking supervision departments focus. November 2005, the Basel Committee launched "New Basel Capital Accord". New Capital Accord puts higher requirements on bank credit risk management and data and quantitative approach. Although Chinese commercial banks have raised the awareness about market risks and the supervision levels supervision departments have also improved, the research and practice on credit, operational and market risk is still in the exploratory stage. Personal credit is the basis of market economy; without personal credit, the market economy will lose solid foundation.
This paper, with bank loan credit risk as the core, mainly studies the performance of different methods of data mining in commercial bank loan credit risk assessment, including neural network algorithm, genetic algorithms and decision tree algorithm. First of all, we use the three algorithms to construct credit risk assessment model separately, and then analyze the loan data with the models we have constructed and get the results of credit risk assessment. Finally, we compare and analyze the test results of the three models, then study the strengths and weaknesses of different algorithm in credit risk assessment and come to the conclusion.
引文
1张光平.从巴塞尔协议看我国金融机构风险监管. www.stockstar.com
2陈卫东.新巴塞尔资本协议评析.国际金融研究. 2001,(3):4~7
3崔滨洲.新巴塞尔协议与中国银行业的全面风险管理.武汉大学学报(哲学社会科学版). 2004,57(3):364~369
4唐文玉.国外个人信用制度的经验及对中国的启示.湖南社会科学. 2001,(6):60~63
5 Kenneth Carling, Tor Jacobson, Kasper Roszbach. Dormancy Risk and Expected Profits of Consumer Loans . Journal of Banking & Finance. 2001,(25):717~739
6喻天辉.商品房屋购买中按揭风险和防范.北京房地产. 2004,(8):86~87
7刘军丽,陈翔.基于决策树的个人住房贷款信用风险评估模型.计算机工程, 2006,32(13):263~265
8 Altman E I. Commercial Bank Lending: Process, Credit Scoring, and Costs of Errors in Lending . Journal of Financial and Quantitative Analysis. 1980,15(4): 813~832
9 Y.E. Orgler. A Credit Scoring Model for Commercial Loans. Journal of Money, Credit and Banking. 1970, l2(4):435~445
10 R. H. Davis, D. B. Edelman, A. J. Gammerman. Machine Learning Algorithms for Credit-Card Applications. IMA Journal of Mathematics Applied in Business and Industry. 1992,(4):43~51
11 Crouhy M, Galai D, MarkR. A Comparative Analysis of Current Credit Risk Models. Journal of Banking & Finance. 2000,(24):59~117
12 Altman E I, Caoutte J B, Narayanan P. Credit Risk Measurement Management: The Ironic Challenge in the Next Decade. Financial Analysis Journal. 1998,(9):7~11
13 Henley, Hand. A Nearest Neighbor Classification for Assessing Consumer Credit Risk. The Statistician. 1996,45(1):1
14孙冰,刘洪玉.信息不对称下的个人住房抵钾货款交易与风险.华中师范大学学报. 2006,(5):63~68
15徐佳娜,西宝.基于AHP-ANN模型的商业银行信用风险评估.哈尔滨理工大学学报. 2005,(5):21~26
16张永娟,张志鸿,霍学喜.信用风险程度模糊综合评价模型.金融教学与研究. 2004(2):13~15
17王春峰,赵欣,韩冬.基于改进蚁群算法的商业银行信用风险评估方法.天津大学学报. 2005,7(2):81~85
18周志翠.浅谈消费信贷中的个人信用评估.邮政研究. 1999,15(4):12~13
19 Tor Jacobson, Kasper Roszbach. Bank Lending Policy, Credit Scoring and Value-at-Risk. Journal of Banking & Finance. 2003,(27):615~633
20 Edward I. Altman, Anthony Saunders. Credit Risk Measurement: Developments over the last 20 Years. Journal of Banking & Finance. 1998,(21):1721~1742
21浅析信用风险评价方法. www.credit.com. 2008
22 Jonathan N.Crook, David B.Edelman, Lyn C. Thomas. Recent Developments in Consumer Credit Risk Assessment. European Journal of Operational Research. 2007,(183):1447~1465
23王建平.数据挖掘技术的发展与研究.情报杂志. 2007,(6):79~80
24杨辉.数据挖掘及其在商业银行中的应用.中国金融电脑. 1998,(11):27~29
25李荣珍.浅谈数据挖掘技术.计算机与网络. 2007,(16):45
26梁剑锋.数据挖掘技术现状和发展.大众科学. 2007(12):2
27 Michael Doumpos·Constantin Zopounidis. Model Combination for Credit Risk Assessment: A Stacked Generalization Approach. Ann Oper Res, 2007,(151): 289~306
28谈恒贵,王文杰,李游华.数据挖掘分类算法综述.微型机与应用. 2005,2:4~9
29姚渝,向伟.信息技术在全球银行业的应用——数据挖掘技术及其应用.中国城市金融. 2001(6):52~54
30张胡,张徽燕.数据挖掘技术在银行业的应用.长春师范学院学报(自然科学版). 2006,25(1):61~64
31王小燕,周建民.数据挖掘技术在商业银行中的应用研究.华南金融电脑. 2005(5):94~96
32于立勇.商业银行信用风险评估预测模型研究.管理科学学报. 2003,6(5):46~53
33 Chih~Fong Tsai, Jhen-Wei Wu. Using Neural Network Ensembles for Bankruptcy Prediction and Credit Scoring. Expert Systems with Applications. 2007,(8):15~21
34 Lean Yu, Shouyang Wang, Kin Keung Lai. Credit Risk Assessment with a Multistage Neural Network Ensemble Learning Approach. Expert Systems with Applications. 2007,(12):11~16
35 J.GALINDO. Credit Risk Assessment Using Statistical and Machine Learning: Basic Methodology and Risk Modeling Applications. Computational Economics. 2000,(15):107~143
36 Yueh-Min Huang, Chun-Min Hung, Hewijin Christine Jiau. Evaluation of Neural Networks and Data Mining Methods on a Credit Assessment Task for Class Imbalance Problem. Nonlinear Analysis: Real World Applications. 2006,(7):720~747
37 THOMAS L C. A Survey of Credit and Behavioral Scoring: Forecasting Financial Risk of Lending to Consumers. International Journal of Forecasting. 2000,16(12):149~172
38 Tam K Y, Kiang M Y. Managerial. Applications of Neural Networks: The Case of Bank Failure Predictions. Management Science. 1992,38(7):926~947
39 W M Spears, K A. De Jong. Using Genetic Algorithms for Supervised Concept Learning. Artificial Intelligence Method sand Applications[C]. World Scientific. 1992,6(5):123~155
40 De Jong K A, Spears W M, Gordon D F. Using Genetic Algorithms for Concept Learning. Machine Learning. 1993,13(23):161~188
41 Cardie C. Using Decision Trees to Improve Case-based Learning [C]. Proceedings of the 10th International Conference on Machine Learning. 1993,(2): 25~32
42 Han Jiawei, Kamber M.数据挖掘概念与技术.范明,孟小峰译.机械工业出版社, 2001:219~240
2陈卫东.新巴塞尔资本协议评析.国际金融研究. 2001,(3):4~7
3崔滨洲.新巴塞尔协议与中国银行业的全面风险管理.武汉大学学报(哲学社会科学版). 2004,57(3):364~369
4唐文玉.国外个人信用制度的经验及对中国的启示.湖南社会科学. 2001,(6):60~63
5 Kenneth Carling, Tor Jacobson, Kasper Roszbach. Dormancy Risk and Expected Profits of Consumer Loans . Journal of Banking & Finance. 2001,(25):717~739
6喻天辉.商品房屋购买中按揭风险和防范.北京房地产. 2004,(8):86~87
7刘军丽,陈翔.基于决策树的个人住房贷款信用风险评估模型.计算机工程, 2006,32(13):263~265
8 Altman E I. Commercial Bank Lending: Process, Credit Scoring, and Costs of Errors in Lending . Journal of Financial and Quantitative Analysis. 1980,15(4): 813~832
9 Y.E. Orgler. A Credit Scoring Model for Commercial Loans. Journal of Money, Credit and Banking. 1970, l2(4):435~445
10 R. H. Davis, D. B. Edelman, A. J. Gammerman. Machine Learning Algorithms for Credit-Card Applications. IMA Journal of Mathematics Applied in Business and Industry. 1992,(4):43~51
11 Crouhy M, Galai D, MarkR. A Comparative Analysis of Current Credit Risk Models. Journal of Banking & Finance. 2000,(24):59~117
12 Altman E I, Caoutte J B, Narayanan P. Credit Risk Measurement Management: The Ironic Challenge in the Next Decade. Financial Analysis Journal. 1998,(9):7~11
13 Henley, Hand. A Nearest Neighbor Classification for Assessing Consumer Credit Risk. The Statistician. 1996,45(1):1
14孙冰,刘洪玉.信息不对称下的个人住房抵钾货款交易与风险.华中师范大学学报. 2006,(5):63~68
15徐佳娜,西宝.基于AHP-ANN模型的商业银行信用风险评估.哈尔滨理工大学学报. 2005,(5):21~26
16张永娟,张志鸿,霍学喜.信用风险程度模糊综合评价模型.金融教学与研究. 2004(2):13~15
17王春峰,赵欣,韩冬.基于改进蚁群算法的商业银行信用风险评估方法.天津大学学报. 2005,7(2):81~85
18周志翠.浅谈消费信贷中的个人信用评估.邮政研究. 1999,15(4):12~13
19 Tor Jacobson, Kasper Roszbach. Bank Lending Policy, Credit Scoring and Value-at-Risk. Journal of Banking & Finance. 2003,(27):615~633
20 Edward I. Altman, Anthony Saunders. Credit Risk Measurement: Developments over the last 20 Years. Journal of Banking & Finance. 1998,(21):1721~1742
21浅析信用风险评价方法. www.credit.com. 2008
22 Jonathan N.Crook, David B.Edelman, Lyn C. Thomas. Recent Developments in Consumer Credit Risk Assessment. European Journal of Operational Research. 2007,(183):1447~1465
23王建平.数据挖掘技术的发展与研究.情报杂志. 2007,(6):79~80
24杨辉.数据挖掘及其在商业银行中的应用.中国金融电脑. 1998,(11):27~29
25李荣珍.浅谈数据挖掘技术.计算机与网络. 2007,(16):45
26梁剑锋.数据挖掘技术现状和发展.大众科学. 2007(12):2
27 Michael Doumpos·Constantin Zopounidis. Model Combination for Credit Risk Assessment: A Stacked Generalization Approach. Ann Oper Res, 2007,(151): 289~306
28谈恒贵,王文杰,李游华.数据挖掘分类算法综述.微型机与应用. 2005,2:4~9
29姚渝,向伟.信息技术在全球银行业的应用——数据挖掘技术及其应用.中国城市金融. 2001(6):52~54
30张胡,张徽燕.数据挖掘技术在银行业的应用.长春师范学院学报(自然科学版). 2006,25(1):61~64
31王小燕,周建民.数据挖掘技术在商业银行中的应用研究.华南金融电脑. 2005(5):94~96
32于立勇.商业银行信用风险评估预测模型研究.管理科学学报. 2003,6(5):46~53
33 Chih~Fong Tsai, Jhen-Wei Wu. Using Neural Network Ensembles for Bankruptcy Prediction and Credit Scoring. Expert Systems with Applications. 2007,(8):15~21
34 Lean Yu, Shouyang Wang, Kin Keung Lai. Credit Risk Assessment with a Multistage Neural Network Ensemble Learning Approach. Expert Systems with Applications. 2007,(12):11~16
35 J.GALINDO. Credit Risk Assessment Using Statistical and Machine Learning: Basic Methodology and Risk Modeling Applications. Computational Economics. 2000,(15):107~143
36 Yueh-Min Huang, Chun-Min Hung, Hewijin Christine Jiau. Evaluation of Neural Networks and Data Mining Methods on a Credit Assessment Task for Class Imbalance Problem. Nonlinear Analysis: Real World Applications. 2006,(7):720~747
37 THOMAS L C. A Survey of Credit and Behavioral Scoring: Forecasting Financial Risk of Lending to Consumers. International Journal of Forecasting. 2000,16(12):149~172
38 Tam K Y, Kiang M Y. Managerial. Applications of Neural Networks: The Case of Bank Failure Predictions. Management Science. 1992,38(7):926~947
39 W M Spears, K A. De Jong. Using Genetic Algorithms for Supervised Concept Learning. Artificial Intelligence Method sand Applications[C]. World Scientific. 1992,6(5):123~155
40 De Jong K A, Spears W M, Gordon D F. Using Genetic Algorithms for Concept Learning. Machine Learning. 1993,13(23):161~188
41 Cardie C. Using Decision Trees to Improve Case-based Learning [C]. Proceedings of the 10th International Conference on Machine Learning. 1993,(2): 25~32
42 Han Jiawei, Kamber M.数据挖掘概念与技术.范明,孟小峰译.机械工业出版社, 2001:219~240