The Design And Implementation Of Commercial Bank Credit Risk Model Based On PCA And BPNN

With social awareness of credit risk and benign asset cycle philosophy,commercial bank must establish an effective credit risk platform for monitoring and control.This paper investigates the relevance analysis method and the commercial bank credit risk assessment model,which can promote our financial industry to realize healthy development,promote our country establishing credit risk assessment system and prevent global credit crisis.Based on the theory of risk and credit risk management,this paper builds a neural network nonlinear model of commercial bank credit risk.First of all,based on theoretical analysis of credit risk,which is to determine the comprehensive evaluation index of credit risks;Secondly,with the analysis of the credit grade evaluation factors in?the customer credit evaluation report?,some key process indicators is determined which may impact credit evaluation;Then,on the basis of mass sample data in the database of commercial bank loan enterprise,a sample set is made;Finally,several learning algorithms of neural networks is compared in performance and error convergence,and LM-BP credit risk assessment model is established,but the prediction precision of the model has yet to be improved.A BP neural network model is established,whose noise and dimensionality are reduced by principal component analysis.The LM-BP neural network and improved PCA-LM neural network simulation are compared.The application of credit risk management model of commercial bank,achieves distinct economic benefit and social benefit.Meanwhile a set of practical,efficient and recommendable industrialized methods is supplied for the credit risk problems.