Ensemble prediction of commercial bank failure through diversification of input features

Authors:
Sung Woo Shin;Kun Chang Lee;Suleyman Bilgin Kilic
Affiliations:
School of Business Administration, Sungkyunkwan University, Seoul, Korea;School of Business Administration, Sungkyunkwan University, Seoul, Korea;Faculty of Economic and Administrative Science, Cukurova University, Balcali, Adana, Turkey
Venue:
AI'06 Proceedings of the 19th Australian joint conference on Artificial Intelligence: advances in Artificial Intelligence
Year:
2006

Citing 9
Cited 1

Learning internal representations by error propagation

Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
C4.5: programs for machine learning

C4.5: programs for machine learning
Managerial applications of neural networks: the case of bank failure predictions

Management Science
Decision Combination in Multiple Classifier Systems

IEEE Transactions on Pattern Analysis and Machine Intelligence
Democracy in neural nets: voting schemes for classification

Neural Networks
Bagging predictors

Machine Learning
Neural Network Ensembles

IEEE Transactions on Pattern Analysis and Machine Intelligence
Application of majority voting to pattern recognition: an analysis of its behavior and performance

IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Bankruptcy prediction for credit risk using neural networks: A survey and new results

IEEE Transactions on Neural Networks

Predictive model of insolvency risk for Australian corporations

AusDM '07 Proceedings of the sixth Australasian conference on Data mining and analytics - Volume 70

Quantified Score

Hi-index	0.00

Visualization

Abstract

As primary focus of banking regulation and supervision is being shifted toward internal risk management for all commercial banks, financial data mining task such as an early warning of bank failure becomes more critical than ever. In this study, we examine the effect of variable selection methods for intelligent bankruptcy prediction models. Moreover, an augmented stacked generalizer that utilizes diversified feature subsets during its learning phase is suggested as an effective ensemble method for promoting independencies among base prediction models. Empirical results show that the augmented stacked generalizer significantly improves overall predictability by reducing the more costly type-I error rate compared against both popular bagging and standard stacking procedures.