The Integrated Methodology of Rough Set Theory and Support Vector Machine for Credit Risk Assessment

Abstract

According to the current situation of the credit risk assessment in commercial banks, a hybrid intelligent system is applied to the study of credit risk assessment in commercial banks, combining rough set approach and support vector machine (SVM). The information table can be reduced, which showed that the number of evaluation criteria such as financial ratios and qualitative variables was reduced with no information loss through rough set approach. And then, the reduced information table is used to develop classification rules and train SVM. The rationality of hybrid system is using rules developed by rough sets and SVM. The former is for an object that matches any of the rules and the latter is for one that does not match any of them. The effectiveness of the methodology was verified by experiments comparing traditional discriminant analysis model and BP neural networks with our approach