A Research On Overdue Behavior Of Post-Loan Users Based On Customer Segmentation

With the rapid development of China's economy,traditional banks and lending companies can not meet the needs of the public loaning,and the borrowing audit threshold is increasing.In recent years China's Internet financial enterprises to flourish,to fill the gap and solve the financing of small and medium-sized enterprises,that coundn't apply for general public loan from a bank credit.However,with the expansion of internet finance companies and a sharp increase of customers,the lack of a completed user management system.Any customer's risk behavior parts of pre-loan,loan,after loan,will affect the normal operation of the entire company,lead capital liquidity reduced,the company's operating costs increased.How to effectively segment the borrowers,prevent customers overdue repayment,reduce the company's bad debt rate,to ensure that the company's normal operation and development becomes more and more important.Borrowing this financial behavior with the development of the Internet,it has both the characteristics of finance and the attributes of the Internet.And the audience is wide.the Internet financial Company has a large number of data accumulation.we can apply to different stages of the customer to take different risk management.For the large base of the internet finance Company,the establishment of the company's user management system,can use the idea of clustering,customer segmentation,screening out high-quality customers,ordinary customers,high contribution customers,loyal customers,as well as inferior customers and other different standards.effectively control of the risk of internet finance companies,and is suitable for the company's user management.This is also an important link to achieve precise marketing and risk management.In the past study,we have only focused on the pre-loan customer's risk management and credit evaluation,the study of the default behavior of user is little.This article mainly study for the post-loan users,based on customer relationship management(CRM).Use the K-means cluster analysis of the user groups.These users mainly includes non-defaulting customers after the loan,and default customers.According to the credit status,personal status,resident area,income level,borrowing,consumption behavior and other dimensions,the customer group is divided into 8 customer groups,providing products,services and marketing models in a targeted way.It mainly uses the RFM model to group,use K-means to cluster,to determines the user's personal value,and provides decision suggestions for the business operators.The decision tree model is used to forecast the default,which provides data supportfor pre-loan risk management and user credit grade,thus reducing the overdue rate of the company's customers and providing data supporting for the risk control.