Based On Data Mining Of Personal Banking Customer Credit Scoring Models

Enterprise credit rating is the foundation of credit risk measurement as the New Basle Capital Accord proposed and which is the important method for credit risk management of the bank. In china, the development of database for enterprise credit rating is at the beginning stage and the qualitative analysis is the main method. The essay focused on analysis for how to create a individual credit rating model for commercial bank. Based on housing loan database, analyze how to implicate data mining method to real individual rating database and try to find some uncover rules. The credit rating model will be established finally.Based on the real individual loan database, supported by data mining skill, we use CART method to analyze the database. Followed by data gathering,extracting,previous processing, we establish the individual credit rating'decision tree model. Furthermore, we get the individual credit rating model. And the model we get is based on 100 percent measuring method, it can help bank to establish the rating formula.From the testing result, the model can forecast high level credit client more exactly. But to the middle and low level credit client, the result is less exactly. It is also a tough task to the credit rating. Received or rejected, this is a question. More information must be collected for the more authentic forecast result.The original model is established by objectively sight. And we edit the model by adding subjectively sight. Layer analysis method is used to establish the model. Through we combined the subjectively and objectively method, we get the finally version.