| Summary: | 碩士 === 國立勤益科技大學 === 工業工程與管理系 === 97 === Since 2008 year, The finance seaquake attacks all industry of global to cause enterprises bankrupt and unemployed. The finance seaquake cause from house inferiority extends credit of America. Actually, The house secondary extend credit of America is finance seaquake inferior cause. The major cause of finance seaquake form Credit Structure Goods to overflow. Credit Card business also Credit goods part. The banks violent competition condition, to provide implement high risk and high profit and easy apply Credit Card to promotion consumer. The banks even to simplify Credit review process to attract consumer. Therefore, The banks credit risk to rise and to lead reimbursement delay of credit card; bad debt and cost of receivables on demand substantially grow. Therefore, The review credit risks appearing important.
By above-mentioned requirement, The study program will collect research references for finance risk and credit management. The case study of programs by credit record database from someone of banks. We will apply CRISP-DM methodology. First, We will collect customer basic background data and payment characteristic variable. Second, Apply Artificial Neural Network of Data Mining forecast customer whether regular consume and payment, and combine Decision Tree to improve and establish customer credit principles. Third, Experimental result value verify and control improve effect. Classify of Data Mining have credit discrimination capability very high for credit card. It should establish accurate credit regulations and forecast model. The bank could find potential key factor and credit regulations of credit card by the model. The model could reduce loss cost form type I and type II errors for credit business. And promote steady and make a profit of credit card business.
|
|---|
