模糊近似支持向量机模型及其在信用风险评估中的应用

姚潇, 余乐安

系统工程理论与实践 ›› 2012 ›› Issue (3) : 549-554.

PDF(505 KB)
PDF(505 KB)
系统工程理论与实践 ›› 2012 ›› Issue (3) : 549-554. DOI: 10.12011/1000-6788(2012)3-549
论文

模糊近似支持向量机模型及其在信用风险评估中的应用

    姚潇1, 余乐安1,2,3
作者信息 +

A fuzzy proximal support vector machine model and its application to credit risk analysis

    YAO Xiao1, YULe-an1,2,3
Author information +
文章历史 +

摘要

支持向量机是近些年兴起的人工智能方法,并在信用风险分析领域得到了广泛应用.为了有效地减小在实证研究中样本的奇异点和噪声对模型的干扰, 在近似支持向量机的基础上, 引入模糊隶属度的思想,提出了模糊近似支持向量机, 使之不仅能保留近似支持向量机原有的优点,同时也可以减小奇异点和噪声对模型的干扰,从而进一步提高了分类判别能力. 为验证模糊近似支持向量机的效果, 利用两个公开的信用数据集进行实证研究. 实证研究结果表明:与其它模型相比,所提出的模糊近似支持向量机能够显著地提高信用风险分类精度,具有较高的实用价值.

Abstract

Support vector machines (SVM), one of the emerging artificial intelligence techniques, has been widely used in credit risk evaluation. In order to reduce the effects of outliers and noise in the datasets effectively, a fuzzy proximal SVM (FPSVM) is proposed based on the proximal SVM (PSVM) and fuzzy memberships theory. Due to the fact that the proposed FPSVM can effectively decrease the disturbance of sample outliers and noise without the loss of the advantages of PSVM, the proposed FPSVM can increase the model performance significantly. For verification and illustration, two public available credit datasets are used to test and compare the performance of different models. The experimental results show the proposed FPSVM model can yield better performances compared with other models listed in this study.

关键词

信用风险评估 / 近似支持向量机 / 模糊隶属度

Key words

credit risk analysis / proximal support vector machine / fuzzy memberships

引用本文

导出引用
姚潇 , 余乐安. 模糊近似支持向量机模型及其在信用风险评估中的应用. 系统工程理论与实践, 2012(3): 549-554 https://doi.org/10.12011/1000-6788(2012)3-549
YAO Xiao , YULe-an. A fuzzy proximal support vector machine model and its application to credit risk analysis. Systems Engineering - Theory & Practice, 2012(3): 549-554 https://doi.org/10.12011/1000-6788(2012)3-549
中图分类号: F830   

参考文献

[1]Fisher R A. The use of multiplemeasurements in taxonomic problems[J]. Annals of Eugenics, 1936(7): 179-188.%期刊著录格式
[2]Altman E I. Financial ratios, discriminant analysis and theprediction of corporate bankruptcy[J]. Journal of Finance, 1968, 23: 589-609.%期刊著录格式
[3]Wiginton J C. A note on the comparison of logit anddiscriminant models of consumer credit behavior[J].Journal of Financial and Quantitative Analysis, 1980, 15: 757-770.%期刊著录格式
[4]Grablowsky B J, Talley W K. Probit and discriminantfunctions for classifying credit applicants: A comparison[J]. Journal of Economics and Business, 1981, 33: 254-261.%期刊著录格式
[5]Steenackers A, Goovaerts M J. A credit scoring model forpersonal loans[J]. Insurance: Mathematics and Economics, 1989(8): 31-34.%期刊著录格式
[6]West D. Neural network credit scoring models[J]. Computer& Operations Research, 2000, 27: 1131-1152.%期刊著录格式
[7]Baesens B, Setiono R, Christophe M, et al. Using neuralnetwork rule extraction and decision tables for credit-riskevaluation[J]. Management Sciences, 2003, 49: 312-329.%期刊著录格式Journal of Finance, 1968, vol.23: 589-609.%期刊著录格式
[8]吴冲, 吕静杰, 潘启树, 等.基于模糊神经网络的商业银行信用风险评估模型研究[J].系统工程理论与实践, 2004, 24(11): 1-8.Wu C, Lü J J, Pan Q S, et al. Study on credit risk assessmentmodel of commercial banks based on fuzzy neural network[J].Systems Engineering - Theory & Practice, 2004, 24(11): 1-8.
[9]王春峰, 万海晖, 张维.基于神经网络技术的商业银行信用风险评估[J]. 系统工程理论与实践,1999, 19(9): 24-33. Wang C F, Wan H H, Zhang W. Credit riskassessment in commercial banks using neural networks[J]. SystemsEngineering - Theory & Practice, 1999, 19(9): 24-33.
[10]Vapnik V. The Nature of Statistical Learning Theory[M].New York: Springer, 1995.
[11]Vapnik V. Statistical Learning Theory[M]. New York: JohnWiley and Sons, 1998.
[12]Suykens J A K, Vandewalle J V. Least squares supportvector machine classifiers[J].Neural Processing Letters, 1999(9): 293-300.%期刊著录格式
[13]Suykens J A K, van Gestel T, De Brabanter J, et al.Least Squares Support Vector Machines[M]. World Scientific,Singapore, 2002.
[14]Fung G, Mangasarian O L. Proximal support vector machineclassifiers[C]// Proceedings of the Seventh ACM SIGKDDInternational Conference of Knowledge Discovery and Data Mining,San Francisco, California, USA, 2001: 77-86.
[15]Lin C F, Wang S D. Fuzzy support vectormachines[J]. IEEE Transactions on Neural Networks, 2002, 13: 464-471.%期刊著录格式
[16]Huang H P, Liu Y H. Fuzzy support vector machines forpattern recognition and data mining[J].International Journal of Fuzzy Systems, 2002(4): 826-835.%期刊著录格式
[17] Abe S, Inoue T. Fuzzy support vector machines formulti-class problems[C]// Proceedings of the Tenth EuropeanSymposium on Artificial Neural Networks, Bruges, Belgium, 2002:113-118.
[18]Hardle W, Moro R A, Schafer D. Predicting bankruptcywith support vector machines[J]. Statistical Tools for Finance and Insurance, 2005: 225-248.%期刊著录格式
[19]van Gestel T, Baesens B, Garcia J, et al. A supportvector machine approach to credit scoring[J].Bank en Financiewezen, 2003(2): 73-82.%期刊著录格式
[20]Baesens B, van Gestel T, Viaene S, et al. Benchmarkingstate-of-art classification algorithms for credit scoring[J].Journal of the Operational Research Society, 2004, 54: 627-635.%期刊著录格式
[21]Yu L, Wang S Y, Cao J. A modified least squares supportvector machine classifier with application to credit riskanalysis[J]. International Journal of Information Technology andDecision Making, 2009(8): 697-710.%期刊著录格式
[22] Yu L, Wang S Y, Lai K K. An intelligent-agent-basedfuzzy group decision making model for financial multicriteriadecision support: The case of credit scoring[J]. European Journal of Operational Research, 2009, 195: 942-959.%期刊著录格式
[23]Yu L, Wang S Y, Wen F H, et al. Designing a hybridintelligent mining system for credit risk evaluation[J]. Journal of Systems Science & Complexity, 2008, 21(4): 527-539.%期刊著录格式
[24]Mercer J. Functions of positive and negative type, andtheir connection with the theory of integral equations[J].Transactions of the London Philosophical Society, 1909, 209: 415-446.%期刊著录格式
[25]Mathworks. Matlab R2008a[CP]. 2008.
[26]Fawcett T. An introduction to ROC analysis[J]. PatternRecognition Letters, 2006, 27: 861-874.

基金

国家杰出青年科学基金(71025005);国家自然科学基金重大研究计划培育项目~(90924024)
PDF(505 KB)

117

Accesses

0

Citation

Detail

段落导航
相关文章

/