Design And Implementation For A Bank Customer Loss Warning System

Under the background of big data,commercial banks have accumulated large amounts of customer data,but how to identify the underlying features of corporate behavior has become an urgent problem to be solved.However,the current customer relationship management bank remained at the subjective level management,such as due diligence,client manager and regular visits,etc.,such approach,while maintaining customer timely capture the customer’s status,but with normalization after retail customers,clients relations gradually rising maintenance costs.Therefore,building a customer can proactively predict stability system has become an important part of the bank A dynamic management of customer relationships.This article is based on this starting point,the use of accumulated A bank customer data to build customer loss warning system,with a view to cost-efficiency of customer status judgment.First of all,this paper combs the construction of customer loss early warning system of the main basic theory.We discuss the basic meaning,main function and basic process of data mining technology,and briefly review the development of data mining technology.Then,this paper focuses on the four popular data mining technologies: decision tree analysis,neural network,genetic algorithm and support vector machine.At last,Spring MVC and My Batis,which are the lightweight architecture technology of Web platform of J2 EE platform,are sorted out.Third,the analysis of the demand for customer churn warning system.First,the analysis of the factors of customer churn,systematically explain the motives of customer churn.Secondly,according to the reasons of custome r churn,the construction demand of customer churn warning system is put forward: system demand and business demand.Fourth,customer churn system architecture and design.Firstly,the goal and principle of system design are expounded.Secondly,the overal l construction of the system is discussed.Thirdly,the design of system database is described.At last,the design of the core module is expounded.Fifth,this paper carries on the measurement to the A bank customer churn warning system.Firstly,the interface of the system is described,and the results of the core module are tested.Secondly,the robustness of the system is tested.Thirdly,the prediction results of the combined system and the single system are compared and analyzed.Finally,the use of customer loss warning system to predict the current data and found that more than the size of the customer is relatively stable,and small micro-customers a higher warning rate.In the end,this paper puts forward a combination of objective evaluation and subjective judgment,adjusts the early warning model timely,establishes the response plan of customer loss feedback,and balances the new customer marketing and control customer churn.