The Application Research Of Association Rules Mining Technology In Management Of Taxation Law Enforcement

Insisting in managing taxation by law, building a justice and justness environment of taxation law execution, is a very important part of building harmonious socialism society, and the start and stand point of current taxation system developing taxation work. How using today's information technology, especially the data mining technology, to analyze the taxation law execution data that we got, and provide useful decision information for taxation system to enhance management of taxation law execution has been a hot research point of taxation system. Based on the deep research of association rules mining technology, this dissertation pointed out the advantages and disadvantages of several classic association rules algorithms, and developed a algorithm that based onhash and item sets space division—SAVM algorithms. It also applyassociation rules technology and SAVM algorithms to analyzing the taxation law execution data, found the relationship between fault actions of taxation law execution and the one who make them, provide some very valuable reference information for the taxation system to enhance tax law execution.The main job and innovation points of this dissertation are:1. Described the current situation of data mining technology applied in taxation system, given a full version of taxation law execution management's content, methods and weak points, analyzing how important to apply data mining to analyzing taxation law execution.2 . Introduced the basic theory, main technology, knowledge classification, common used tools and the research trend. Deeply researched association rules technology, and introduced some classical association rules algorithms by stress. Analyzed the advantages and disadvantages of all algorithms, and pointed out the research hot point and develop trend of association rules mining technology.3. Developed a algorithm that based on hash and item sets spacedivision—SAVM algorithms. It converted transaction database into bitevector, by the vector's orperation to get item sets support degree, and just one time scan in database. Within hash technology, it build frequent 2 item sets by one time scan database, to improve the efficiency of algorithm. By using the common prefix, it divided frequent item sets into different sub spaces. Candidate item sets were built in every independent sub space to decrease the connection times. The experiments result tells us SAVM algorithms improved the efficiency of association rules mining.4. Apply data mining technology to taxation law execution data analyzing, explained the data choose, preprocess, mining frequent item sets and develop association rules in detail. By analyzing the developed rules, it found the valuable reference information for taxation system enhance tax law execution. At the same time, it supplied a useful reference method for applying association rules mining technology to other fields.