Research And Application Of Self-learning Warning System Based On Business Rules

When business needs change, it is very difficult for the current audit early warningsystem to update the relevant business logic, and the system itself does not have theself-learning ability in business rules, cannot discover new business rules and lack of abilityto improve their own set of business rules. Based on the theory of business rules, combinedwith Data Mining techniques, a research on the self-learning early warning system basedon business rules is proposed to meet the audit warnings’ real-time requirements for theupdates of the business rules and requirements of the self-learning ability in business rules.Firstly, the relative theory of the business rules is studied in depth, and the keytechniques of business rules, the methodology of business rule and business rule managementare elaborated in this paper. Secondly, the audit business rule repository is built on the basisof theoretical studies on business rules, and the details of specific work for building, whichincludes the collection of audit business logic, the design of the business object model and theformulation of the business rules set, is given. Then, on the basis of researchresults from home and abroad, the early warning model structure of self-learning is designed,and the module functions, working principles and model framework of it is studied in depthfrom the audit early warning system’s requirements of updating the business rules in real-timeand self-learning. Based on the knowledge base theory, the representation and storage formsof the policy of knowledge are given, and the external trigger event and the event priority aredefined in the early warning mechanism; The clustering algorithm of KNNFC applying tothe audit field is proposed, and the experimental comparison to the audit data clustering withother clustering algorithms is given in the self-learning mechanism; The rule decompositionand rule mapping algorithm, which successfully parses the business execution language, isgiven in the language.Finally, on the basis of the aforementioned study, a self-learning audit early warningsystem based on business rules is designed and realized, the design and developmentprograms of the system is given, the development environment and deployment environmentof the system is introduced briefly, and the feasibility of the technology research is verifiedthrough relative experiments in this paper.