Research On Personal Credit Risk Evaluation System In Banking Based On Combined Forecast Model

Along with the rapid development of China's economy, the scale of individual credit business has expanded rapidly, and various kinds of personal consumption loans including housing mortgages, auto loans, education loans, credit card, etc. has become the new profit growth point of our country commercial banking. But relative to the enterprise credit risk assessment, the domestic banking on credit risk management method is relatively backward. The bank staffs mainly rely on the subjective judgment to decide whether approve loan applicants. They do not have a mature steady model to auxiliary evaluation. So the individual desire always influences evaluation results. At the same time, the long time for examination and approval, the low efficiency and high cost, seriously hindered the development of individual consumer loan business.Personal credit evaluation in banking is the process for evaluating the degree of consumers'credit according to the client's personal information. Credit evaluation is the most important one of the core management technology for entities; therefore, it is necessary to study the model and method of customer credit evaluation banking.This article has received real consumer loan samples in commercial bank from the Zhengzhou, based on which, the article analyzes the assessment mode and method on personal credit risk and constructs the personal credit risk assessment system by studying the characteristics of individual consumer credit business based on data mining methods. Main work include: analyze , select and establish the alternative index collection of commercial personal credit risk assessment system suited to China's national conditions by learning the domestic and foreign existing reference; according to the actual data in our commercial banking system, determine the evaluation index ;combine Logistic regression and RBF neural network method to establish a linear combination forecast model on personal credit evaluation; analyze and inspect the established personal credit evaluation model, then contrast the accuracy and robustness with single forecasting model. Due to the statistical model has better accuracy and the neural network model has better robustness, so the combined forecasting model has both good accuracy and good robustnessThis topic's research and results are as follows:1) Establish a index collection for personal credit risk assessment system suitable for China's national conditions;2) Clean and transform the data in the database;3) choose the single model of risk assessment: Logistic regression method and RBF neural network method;4) Combine two kinds of single method, and establish the combined credit risk assessment model;5) Train and exam the model using sample data.The result of this paper is helpful for promoting commercial bank personal credit risk assessment system construction in our country, and for improving individual consumer credit business in commercial bank. It also can improve the efficiency of examination, the customer satisfaction, and reduce bad assets in banking. It has great practical significance in promoting the development of the consumer credit market and improving the core competitiveness of commercial Banks in China.