A multiple intelligent agent system for credit risk prediction via an optimization of localized generalization error with diversity

Abstract

Company bankruptcies cost billions of dollars in losses to banks each year. Thus credit risk prediction is a critical part of a bank’s loan approval decision process. Traditional financial models for credit risk prediction are no longer adequate for describing today’s complex relationship between the financial health and potential bankruptcy of a company. In this work, a multiple classifier system (embedded in a multiple intelligent agent system) is proposed to predict the financial health of a company. In our model, each individual agent (classifier) makes a prediction on the likelihood of credit risk based on only partial information of the company. Each of the agents is an expert, but has limited knowledge (represented by features) about the company. The decisions of all agents are combined together to form a final credit risk prediction. Experiments show that our model out-performs other existing methods using the benchmarking Compustat American Corporations dataset.

Keywords

Credit rating / business intelligence / localized generalization error / multiple classifier system / feature grouping

References

[1] Altman E.. Bankruptcy, Credit Risk, and High Yield Junk Bonds, 2001, Massachusetts: Blackwell Publishers

[2] Beaver W.H., McNichols M.F., Rhie J.W.. Have financial statements become less informative? Evidence from the ability of financial ratios to predict bankruptcy. Review of Accounting Studies, 2005, 10(1): 93-122.

[3] Breiman L.. Bagging predictors. Machine Learning, 1996, 24(2): 123-140.

[4] Chan, A.P.F., Chan, P.P.K., Ng,W.W.Y., Tsang, E.C.C. & Yeung, D.S. (2007). A novel feature grouping method for ensemble neural network nsing localized generalization error model. Submitted to International Journal on Pattern Recognition and Artificial Intelligence

[5] Chan, A.P.F., Ng, W.W.Y., Yeung, D.S. & Tsang, E.C.C. (2005a). Multiple classifier system with feature grouping for intrusion detection: mutual information approach. Lecture Notes in Artificial Intelligence, 3683: 141–148

[6] Chan, A.P.F., Yeung, D.S., Tsang, E.C.C. & Ng, W.W.Y. (2005b). Empirical study on fusion methods using ensemble of RBFNN for network intrusion detection. Lecture Notes in Artificial Intelligence, 3930: 682–690

[7] Duffie D., Singleton K.. Credit Risk: Pricing, Measurement and Management, 2003, Princeton: Princeton University Press

[8] Giacinto G., Roli F.. Design of effective neural network ensembles for image classification processes. Image Vision and Computing Journal, 2001, 19(9/10): 699-707.

[9] Ho T.K.. The random subspace method for constructing decision forest. IEEE Transaction on Pattern Analysis and Machine Intelligence, 1998, 20(8): 832-844.

[10] Kuncheva L.I.. Combining Pattern Classifiers: Methods and Algorithms, 2004, New Jersey: Wiley.

[11] Kuncheva L.I.. Special issue on diversity in multiple classifier systems. Information Fusion, 2005, 6(1): 3-4.

[12] Latinne, P., Debeir, O. & Decaetecker, C. (2000). Different ways of weakening decision trees and their impact on classification accuracy of DT combination. Lecture Notes in Computer Science, 1857: 200–209

[13] Ng W.W.Y., Dorado A., Yeung D.S., Pedrycz W., Izquierdo E.. Image classification with the use of radial basis function neural networks and the minimization of localized generalization error. Pattern Recognition, 2007, 40(1): 19-32.

[14] Ng, W.W.Y., Yeung, D.S., Firth, M., Tsang E.C.C. & Wang, X-Z. (2007b). Feature selection using localized generalization error for supervised classification problems for RBFNN. Submitted to Pattern Recognition

[15] Optiz, D.W. (1999). Feature selection for ensembles. AAAI, 379–384

[16] Oza, N.C., & Tumer, K. (2001). Input decimation ensembles: decorrelation through dimensionality reduction. Lecture Notes in Computer Sciences, 2096: 238–247

[17] Poon W., Firth M.. Are unsolicited ratings lower? International evidence from bank ratings. Journal of Business Finance and Accounting, 2005, 32(9/10): 1741-1771.

[18] Tsymbal A., Pechenizkiy M., Cunningham P.. Diversity in search strategies for ensemble feature selection. Information Fusion, 2005, 6(1): 83-98.

[19] Tumer K., Ghosh J.. Analysis of decision boundaries in linearly combined neural classifiers. Pattern Recognition, 1996, 29(2): 341-348.

[20] Yeung, D.S., Ng, W.W.Y., Wang, D., Tsang, E.C.C. & Wang, X-Z. (2007). Localized generalization error and its application to architecture selection for radial basis function neural network. To appear in IEEE Transaction on Neural Networks

[21] Yu E., Cho S.. Ensemble based on GA wrapper feature selection. Computers and Industrial Engineering, 2006, 51(1): 111-116.

[22] Yule G.. On the association of attributes in statistics. Philosophical Transactions of the Royal Society of London, Series A, 1900, 194: 257-319.

[23] Gaspary L.P., Sanchez R.N., Antunes D.W., Meneghetti E.. A SNMP-based platform for distributed stateful intrusion detection in enterprise networks. IEEE Journal on Selected Areas in Communications, 2005, 23(10): 1973-1982.

[24] Olfati-Saber R., Fax J.A., Murray M.. Consensus and cooperation in networked multi-agent systems. IEEE Proceedings, 2007, 95(1): 215-233.