Credit risk evaluation with fuzzy neural networks on listed corporations of China

Neural networks (NNs) have been widely used to evaluate credit risk because of their excellent performances of treating non-linear data with learning capability. However, the shortcoming of neural networks is also significant due to a "black box" syndrome and the difficulty in dealing with qualitative information, which limited its applications in practice. To overcome these drawbacks of NNs, in this study we suggested an adaptive network-based fuzzy inference system (ANFIS), a kind of fuzzy neural network models, to evaluate credit risk on the Chinese listed corporations. The results of this study indicate that the predictive accuracies of ANFIS model are much better than NNs model. An illustrative example is given for demonstration.