Credit Card Applications Based on Data Mining

This paper discusses the credit card application process based on data mining technology using principal component analysis (PCA) as an alternative to traditional methods to do variable selecting: weight of evidence(WOE) and information value(IV). Banks often transform historical data to credit scoring models and divide them into good or bad attribute to decide whether to accept credit from a customer. Data usually include personality, capital, collateral, capacity. For credit card applicants, banks use credit scoring or predictive models to categories applicants and predict their future behaviour, which allows them to determine the probability of defaulting in a specific time frame in the future and thus optimizing model is particularly important.