Design And Implementation Of Bank Customer Risk Analysis System Based On User Behavior

With the rapid development of economic globalization,especially the pace of financial globalization,China's establishment of commercial banks and the liberalization of interest rates have continued to open.Therefore,the competition among commercial banks in China has also intensified.How to maintain and The relationship between customers and grasping the financial risks brought by customers is the hot spot of current information technology application.Therefore,in order to make banks more competitive in the current market,this paper designs and implements a bank customer risk analysis system based on user behavior.This paper introduces the problems faced by bank customer relationships and customer risk development,and explains the domestic and international status of bank customer relationships and customer risks.In the system requirements analysis chapter,the application example diagram shows the user's functional requirements for the system,and analyzes the non-functional requirements of the user from the aspects of security,interface beauty and simplicity,and thus determines that the functional module of the system has bank customers.Information management,bank customer credit collection,bank customer classification,bank customer credit risk assessment,bank customer churn risk warning,bank customer risk program implementation.In the system design chapter,the system is designed through functional structure diagram and database design diagram,and the system architecture is built by Spring MVC framework.Based on the data mining technology of big data in the user behavior risk assessment model design part,the user behavior data is collected and mined,and the user behavior risk assessment model is constructed.Finally,the system is tested for functional and performance.The functional test includes customer information query,credit information export,customer classification management,risk assessment result statistics,and loss risk prediction index management.The performance test is aimed at The browsing of each functional module,the test results reached expectations.Based on the multi-index fusion m RMR feature selection algorithm of user behavior,this paper analyzes the characteristics of customer data.Based on the consideration of the unbalanced characteristics of customer churn data,an unbalanced data processing algorithm based on boundary mixed sampling is proposed to realizethe boundary.The minority class samples in the region and the majority class samples in the non-boundary region are processed,and the data set equalization processing is implemented on the basis of maintaining the original data distribution characteristics,thereby improving the performance of the unbalanced data classification.Then the MIF-m RMR feature selection algorithm is used to select the feature.At the same time,the BMS algorithm proposed in this paper is used to equalize the subset after feature selection.Finally,the equalized data is sent to the classifier model to obtain the classification prediction results,so as to achieve the bank loss customer risk warning.