The Research And Application Of Rule_Apriori And SBVR

Data mining is the process of discovering potential and useful knowledge in large amounts of data. As the main research content of data mining, association rule mining is to find the hidden links or relevance from a large number of data items. Apriori algorithm is the classical algorithm for association rule mining. However, it can only resolve the problem of specific domain. For this narrow application the paper proposed a new rule based Apriori algorithm, which greatly expanded the scope of application.With the growing number of businesses, the traditional configuration of business rules exposes its defects including technical staff oriented configuration style, difficult maintaining, and short life-cycle. The paper presented a business-oriented rules configuration model based on SBVR (Semantics of Business Vocabularies and Business Rules)Therefore, this thesis focuses on the rule based Apriori algorithm and SBVR based business rules technology, mainly including the following aspects:1. Studied the steps of mining association rules. For the narrow application of traditional algorithm, the paper proposed a rule-based Apriori algorithm, named Rule_Apriori. The algorithm added a rule layer based on the original algorithm, which can guide and control the implementation of the Apriori algorithm according to the configuration of the rules.2. Analyzed the shortcomings of traditional business rules technology, and put forward a SBVR based business rules configuration model, which is business oriented and configurable, including three modules:business rule customization, basic services, and background services.3. Applied SBVR-based business rules technology to billing configuration system, and thus verified the flexibility of business rules based on SBVR.4. Applied the Rule_Apriori algorithm to the telecommunication packages and the service of persuading users to switch packages. The telecommunication packages mined association rules from the users order, found the user-friendly packages, so as to provide decision support when customizing packages. The service of persuading users to switch packages found more appropriate packages for users, according to the details records of users' consumption, and so as to make package recommendations for the customers.