| With China’s reform and opening develops deeply, Chinese consumingconception renovated gradually as well as booming of personal credit in the bankingindustry.Personal credit has become the core product of commercial banks in thebusiness.So the analysis of personal credit the risk control technology and highefficiency is of great practical significance for resolving commercial bank personalcredit business risk have.Based on the analysis of all the risks of the personal credit, this paper believesthat the credit risk is the most important risk that commercial banks are facing,moreover the cause of it is analyzed. Through the depth study of credit risk，customercredit scoring，Logistic regression and neural network，in the light of actual situationof credit risk management in our commercial banks, the paper proposed a ALB CreditScoring model which is a combination of neural network and Logistic regressi on formore accurate and scientific decision．Making of banks and agencies below is mymain work：Firstly, by using the AHP theory to obtain user feature weights to find out thefactors important, these important factors were analyzed by Logistic regressionanalysis and BP neural network. Secondly, to analyze the important characteristicvalues by BP neural network and Logistic regression. Lastly a comprehensiveevaluation index of the last user value to decide whether to issue a personal credit ornot.With the registered personal credit data from some commercial banks as the input,then the prediction and stability of the model is tested.The experimental results show that the prediction accuracy of ALB model isbetter than using a single model. The stability of ALB model is more than using theBP model with high efficiency. As the same time the ALB model have a more clearexplanation of significance of variables in the ALB model which can better guide thedevelopment of the approval of the credit policy as well as target customersaccurately direct marketing personnel. |
PDF Full Text Request