Credit Evaluation Model and Application Based on Fuzzy Neural Network

The research establishes a credit evaluation model based on fuzzy neural network. It is used to do two patterns classification on the 106 listed companies of China in 2000. It selects four primary financial indexes: earning per share, net asset value per share, return on equity, and cash flow per share. By analyzing the statistical quantities of every variable of both the training samples and the testing samples, after eliminating 22 abnormal samples, and then only analyzing 84 normal samples. The simulation results show that the credit evaluation model based on fuzzy neural network has high discriminate accuracy rate to those rest normal samples. There is only one misjudge sample. The identification accuracy rate is 98.81%. The research shows that, as a method discussion, the fuzzy neural network algorithm is still worthy to do deep research.