Design And Implementation Of Real-time Analysis System For Banking Supervision Data

China's banking industry is an industry with high risk.It is the basic duty of the People's Bank of China to reduce various financial risks and realize financial stability through effective supervision of the banking industry.The People's Bank of China has a huge amount of regulatory data.Mining and analyzing the regulatory data can help the central bank to understand the current financial situation and make reasonable decisions.The huge volume of regulatory data contains huge value information,and its information value will gradually decline with the passage of time.The timeliness of the central bank's processing and analysis of regulatory data determines its regulatory efficiency and effectiveness.However,with the continuous accumulation of data and the increasing scale of regulatory data,banks have increasingly high requirements for the speed of data processing and analysis,and even hope to get the data analysis results before the data flow into the business system.At present,a provincial branch of the People's Bank of China mainly has the following problems in the analysis and utilization of regulatory data: the data collection process is complex,with serious lag;Slow update of data processing technology and lack of real-time data processing capability lead to high data delivery delay;The data analysis and modeling method is simple,inefficient and lacking in real-time,which makes it impossible to make timely and accurate judgments according to the latest data.In view of the above problems and combined with the requirements of a provincial branch of the People's Bank of China for real-time analysis and processing of regulatory data,this paper aims to study and design a set of real-time analysis system based on regulatory data,and solve the real-time collection and storage of regulatory data by integrating distributed log collection and message middleware technology into the system.The real-time computing technology is used to construct the data processing process to complete the real-time pre-processing of the regulatory data.At last,the realtime analysis model is built by the stream data mining technology to realize the realtime analysis of the regulatory data.The main research work of this paper is as follows:(1)Integrate Logstash,Kafka,and Kettle real-time stream processing tools.In view of the lack of real-time data processing capability in a central branch supervision system,Logstash+Kafka is integrated with coding to achieve real-time collection and storage of supervision data,and Kettle+Kafka is integrated to achieve real-time preprocessing of supervision data.(2)A real-time identification model of suspicious transactions based on Concept Very Fast Decision Tree(CVFDT)algorithm is designed.Based on the demand for realtime analysis of anti-money laundering data of a central sub-branch of anti-money laundering supervision department,this paper studies the flow data mining technology,and finally uses CVFDT algorithm based on decision tree idea to build a real-time suspicious transaction identification model.At the same time,Java is used to realize CVFDT algorithm and integrated into the system.(3)Design and implement the real-time analysis system of bank supervision data.The design and development of the system follow the related framework of software engineering,with Java as the core development language and Spring Boot framework as the development foundation.The implementation of the system uses B/S architecture,the use of front-end separation development pattern,and follows the service-oriented development idea.The system mainly consists of three modules,which respectively correspond to real-time collection of regulatory data,real-time processing of regulatory data and real-time analysis of regulatory data.Through regulatory data used in this system the actual test,the system can provide supervision of real-time data acquisition,real-time processing and real-time analysis capabilities,improve the city center branch of regulatory data using the timeliness and effectiveness,and help the city center branch improve the efficiency of supervision,so as to timely and accurate make regulatory decisions.