Research Of Personal Credit Evaluation Model Based On Neural Network

Credit risk is the primary risk that the banks of our country face with, because of the lack of credit for the loan enterprise and personal, cause the fast increasement of non-performing assets for banks. It is one of the most familiar reasons bank bankrupt. With the ratio personal credit has in bank loan raising, it is fundamental to supply a scientific credit evaluation method for the loan personal to avoid the regeneration of non-performing assets of commercial banks.On the foundation of consider computation quantity, make up reasonable Index System of personal credit evaluation, and obtain related data. In practical economic life, lots of factors can affect personal credit, the relationship between these factors and credit is nonlinear. For the traditional credit evaluation model, can hardly identify the mode that affect credit. The technique of neural network is a course of nature Nonlinear Modeling, has the ability discovering the regulation from a great deal of complicated data. So make use of the LVQ technique to build up personal credit evaluation model, divide the credit grade of the customer, identify customer type, thus certain credit granting quota. Adopting BP network to get the customer credit value of prediction, has the advantage to make a precise judgment for the customer placed in the boundary. Because of shortage of BP network, the adoption of genetic algorithms can optimize it's weights and threshold. Combining global searching ability of GA and local searching ability of BP, so as to establish GA-BP network model, estimate customer credit circumstance. The two kinds of model can use independently, also can make complement for each other.Combine the concrete problem to design the model. In recent years, MATLAB has...