The Application Of Statistical Methods On Consumer Credit Scoring System

In recent years, domestic commercial banks had made the great improvement of the personal credit business. Although some of them have realized the importance of the establishment of consumer credit scoring system, most approval authority is still controlled by the Commissioner of risk management. Since subjective assumptions can not be avoided, more and more loan consulting firms make false appearance for borrowers to get a loan from the bank, these factors are likely to cause a great risk to the bank’s credit business.This article will introduce the main risk point of the personal credit business of commercial bank, and use the information of loan applicants obtained by banks to choose and predict reasonably. By establishing reasonable mathematical models and using statistical methods, personal information will be provided by a certain degree of quantization by means of some statistical software, which could exclude the human factors in the approval process and improve the efficiency of the approval and reduce the risk that banks faced.The experiences were compared between foreign and domestic banks except the data analysis in order to avoid producing simple models. I also analyze the current situation of personal loan in China, using the logistic regression model and decision tree model in personal credit system modeling of commercial banks. I will give a comparison among different models, parameter estimation, test results and get the conclusion. Contact between the results and the fact that the experience and data model combined with detailed analysis of the actual situation. Finally, I will do a little in-depth exploration of how to create a complete consumer credit scoring system.