Professional Risk-control System Of Security Companies Based On Algebra Of Abductive

It's been a star-crossed time since 2001 for Chinese securities market. On June 6, 2005, Shanghai securities index fell down below 1000, billions of treasure in the market was gone. Dozens of securities companies such as Anshan, Dalian, Dapeng Securities, South Security etc. quitted the market one after another. To strengthen the risk management and interior control of securities companies arouse remarkable concern from supervisors and across the industry and every aspect of the market has fully recognized that robust risk management system, complete risk control policies are necessary to promote regularly, robust, efficient and ordered development of securities industry. The vital element for efficient risk control and management system is to establish a good risk prediction and analysis inside a securities company with the support of advanced information technology, to make real-time asset and risk perdition and avoid risks in advance. This thesis presents work in following areas for the design of risk control and analysis in a securities company.Firstly, the construction and technologies of risk management systems from domestic and abroad companies are compared. Then, application of expert systems are analyzed. At last, the eligibility and necessity to use expert system theory in the design of securities risk management and analysis system is introduce.Secondly, Knowledge presentation methods are described. As production system resembles the way of human reasoning with natural presentation and its rule database can decompose into modules and suitable for inference, so it is used in the designing of securities risk control knowledge database. For the undetermined characteristic of securities market, the thesis builds a 'risk model' from the belief algebra in fuzzy inference and established a evaluation and quantification system. It also studies the inference control strategy of expert systems, considering the real-time dealing and huge amount of data and shows that users following forward inference control strategy can provide information. As human beings can give quick responses to input facts,efficient judgment and analysis can be given according to the type and size of securities risks based on previous risk model.Thirdly,Further optimizes the system thorough database tuning, minimizes the number of tables, decomposes tables with connection operation and carefully designed their structure for better performance. The issue is systematically studied.Fourthly The system is based on J2EE enterprise application infrastructure, using XML, Web Service technologies and designed an expert system for securities risk prediction and analysis. Under the surveillance of this system, risky operations during the securities business process and information system is dynamically and statically controlled, so as to achieved the goal of timed and efficient risk management.