Store structure and query operation for classification rule

The necessity of creating a rule management system in data mining was analyzed and a constructing method in a classification rule management system was suggested. Different applications of rules discovered during data mining were discussed. The store structure and query operation of the classification rules were designed according to the abstraction of the results of mining and metadata knowledge. The abstraction also includes the stores of data source, rules discovered, and the operations of creating, appending, removing and finding. Moreover, the applications based on the system were implemented. Compared with related other research work, the characteristic of the system is storing not only results mined but also relative data in mining process, to be more convenient in classification and prediction.