Application And Comparison Of Data Mining In The Classification Of Credit Card User

In recent years,the credit card business has developed rapidly in China,bringing huge business opportunities to the banking.Opportunity and crisis coexist,It not only brings business opportunities to the banking but also the challenges.The information technology develops fast in China,Some finance companies get a lot of credit card users’ data,we must to consider how to use and analyze these data to make a right strategic decision for companies.The traditional credit card rating has been unable to meet the needs of banks.So the data mining technology and the statistics knowledge have become importantly for the bank credit rating industry in the future.This paper will use data mining technology and some statistical knowledge to analyze credit card customers in Taiwan’s Banks to reduce the default rate of credit cards.In this paper,I will use the technology of Decision Trees Classification and Support Vector Machine to decide whether the customer has the violations.For the Decision tree classification method,C5.0 algorithm is used,compared with other decision tree classification algorithm,it will sample concentration of different types of sample quantity for consideration,using information gain rate as a branch node selection criteria.Another kind of support vector machine(SVM)algorithm,using the theory of VC dimension and structural risk minimization principle,compared with the traditional machine learning,and consider not only the structure risk minimization by empirical risk,but also consider the risk of confidence.Finally,Comparing the two methods used in this paper,and giving the basis for judging the method of using in different situations.