The Research Of Credit-rating System Of Small-and-Medium-Sized Enterprises In Commercial Banks

International credit rating system has gradually come to a suit of perfect theories and methods by its long time development. Credit risk analysis has been transformed from the method of finance rating grade and subjective estimation to mainly dynamic econometrics analysis based on capital market theory and the technology of cyber-information disposal. The credit evaluating approach of Chinese commercial bank is still on the phase of traditional finance rating analysis. As to Small-and-Medium-Sized (SMS) enterprises, there is not a credit rating system especially for them, but referring to the system of big corporations. Thus, the level of the SMS enterprises can not be evaluated objectively, which restricts the tendency of the commercial bank's extending credit to them. So it appears the financing difficulty for the SMS enterprises.This thesis tries to start with the choosing for credit rating approach of SMS enterprises, and builds a rating model for them by contrasting with corresponding corporation credit rating approaches, using the experience of foreign advanced ranking organ for references. The model is built with the core of optimized design for the credit rating index system and based on the three-lay BP neural network, which could exert neural network's super ability of self-study and non-linearity disposal. This thesis weakens the impact of artificial factors when confirming the importance coefficients by the training of the credit data of 30 corporations, and validates the accuracy of BP neural network by V-fold Cross-Validation, which is a try of building a credit rating system for the SMS enterprises.