Research On Willingness Of Bank Credit Card Installments

With the developing of China's economy,consumer finance is booming,and the emerging financial product market has widely spread,also with the improvement and changes in people's living standards and consumption concepts,bank credit card installment has become a popular consumption method nowadays.For a large group of credit card holders,effectively locking customers who require credit card installment services can be largely improve the performance of marketing efficiency.Data mining and machine learning has been become important tools for finding such customers.The article uses bank customer's credit card information as case data.Descriptive analysis of the data from three aspects of customer information,credit card information,and transaction information is exploited to classify the characteristics from a qualitative perspective,and then the WOE code and IV information Value are for feature construction and selection.Then an hybird model is established using decision tree and ensemble learning algorithms and optimal parameter selection methods.The evaluation model is also discussed to compare in order to obtain the optimal model.Combined with imbalanced data processing methods,oversampling and undersampling methods,the optional model is improved to obtain more effective results.Customers who are willing to accept credit card installment services are found out.Then the marketing can be foced to the small part of customers to reduce costs,improve marketing success rate and the competitiveness.