Research On Weighted Association Rules

With the advent of the information age, the amount of data in databases increases dramatically. In the face of huge data resources, it is an urgent need for a powerful tool to mine useful knowledge. Data Mining is the new technology to extract useful knowledge from large datum, which is also called database knowledge discovery (DKD) that is the high-level process to extract the unknown, novel, valid and potentially useful knowledge or model from a large amount of data.Association Rules Mining is one of the most important subjects about research and application in the field of Data Mining. Apriori Algorithm is the most basic and the key algorithm for Association Rules Mining. However, the mining rules have some limitations because Apriori algorithm only takes into account the item frequency not the importance of the item property and records in database. To the problem, the dissertation mostly studies weighted association rules and presents the new weighted association rule model and algorithm which is the combination of the horizontal weight and vertical weight. The main contents and results of this dissertation are as follows,1. Introducing the correlative concept of data mining and development, studying the process, technology and application of data mining, and discussing the future development.2. Studying the Association Rules, discussing Apriori algorithm, and analyzing the FP-growth Algorithm of non-frequent itemsets.3. Studying weighted association rules in comprehensive manner, the special in-depth study on Weighted Association Rules Algorithm-MINWAL(O) and the analysis for normalization approaches of horizontal weight.4. Presenting the improved weighted association rule algorithm-MWAL algorithm of the combination of the horizontal weight and vertical weight, and expatiating the realization process of the algorithm.