Research resource-optimizing network and its application in credit rating
Wenxing Bao · Journal of Sichuan University · 2008
To overcome disadvantages of Resource-Assigning Network algorithm such as bad robustness and weak generalization,Resource-Optimizing Network is proposed,which is combined sliding data window with grads.Meanwhile,to avoid numerous and jumbled network,this algorithm achieves combination of hidden neurons.On the basis,a credit rating model based on mixed neural network is proposed and applied to lots of listed companies.The result shows the feasibility and effectiveness of the evaluation model.