The Design Of Banking Anti-fraud System Based On Dimensional Modeling

The distance between the speed of Chinese financial market opening up and the speed of Chinese financial market management is becoming farther and farther when it comes to the 21 st century. This makes more and more frauds. In order to monitor and prevent all kinds of banking risks, a design of banking anti-fraud system based on dimensional modeling on banking data warehouse is put forward. The infrastructure of system is DW/BI. The number of anti-fraud models can be extensible on the system.First introduces the architecture of banking anti-fraud system and specifies the theory of dimensional modeling briefly. Then gives the scope of research. The contents of research are divided into three parts. The first part is dimensional modeling on the basic layer of DW aimed at anti-fraud application. Monitoring the banking risks on the banking data warehouse is a kind of data analysis. Dimensional modeling is aimed at data analysis, so it fits the banking anti-fraud system well. The second part is the design of anti-fraud model interface on the market layer of DW and how to implement the business logic using dimensional modeling. Each anti-fraud application is refined to an anti-fraud model. The model includes input and output information and the core business logic. It is convenient to implement the complex business logic with dimensional modeling. The last part is the design of banking anti-fraud BI system. The functions of the system include the anti-fraud model scheduling, the model parameter configuration, the display of model output and so on. The system can load any anti-fraud model which conforms to the model interface according to the needs.Finally it gives the application effect according to the experiment of using the system in one big commercial bank. It compares five quotas including the detection rate of frauds before and after using the system, the frequency of model running before and after using the system and so on based on the same banking data warehouse. The application effect indicates that the system can help banks inspect the risks of fraud effectively.