Credit Card Application Process And Customer Segmentation Model Study

The global economic crisis occurred in 2008, began in the United States subprime mortgage crisis and gradually spread to the real economy with Lehman Brothers and a number of the centuries-old investment bank collapse, financial crisis become the economic crisis, give the global economy a disastrous effect. People unemployed, mortgages and all kinds of credit card loaned have become doubtful and bad debts, The more developed credit, the harm is more serious. In recent years, with China's rapid economic development, the credit card business in China has been an unprecedented golden period of development, for each commercial bank has brought huge resources and considerable profits from customers. However, making a highly competitive review of all applications for credit card issuing bank for people who tend to loose restrictions, ignoring the risk management and control.Through careful analysis and study of credit card holder information, I presented how to control operational risks of credit card, considering two ways:First, through credit card customer segmentation, identify quality customers, star customers and the public risk customers. For each sub-group, the banks should develop corresponding marketing and maintenance strategies, effective use of bank resources, reduce business risk.Second, establish an intelligent credit card application process model, from the large number of credit card applicants quickly and accurately elect to meet the requirements of the customer payment card.Therefore, corresponding to these, this paper propose the customer segmentation models and model credit card application process, and using an open credit card data of a German company to do an empirical analysis. This is the first to Particle Swarm Optimization based on BP Neural Network Model in credit card application process, and compare to the traditional BP neural network model, the results showed that the traditional BP neural network model was soon trapped in local minimum. Therefore, whether the data for training or test data, their accuracy is much lower than BP neural network based on PSO algorithm model. And customer segmentation model is based on principal component analysis and cluster analysis model, eventually broken down into four kinds of credit card customers, and to give the corresponding customer marketing and maintenance.