Association Rules And Incremental Updating Of Association Rules

Along with the rapid development and wide application of information technology, there are so much data stored in database in kinds of fields. The basic requirement to transform unorganized data to information or knowledge is so important that forces the data mining technology becomes one of the most active analysable directions. More than that, in this field, the mining technique based on association rule is full of practical potential.Based on the research of association rule mining, this thesis introduced the method of association rule mining and the increment updating technology. So far the research of incremental updating about the association rule mining focused on the positive association rule mining, where are some classical algorithms such as FUP and FUP2 supported by D.W.Cheung and IUA and PIUA supported by Feng Yucai. However, there lacks literature about the negative association rule mining. By the knowledge of the association rule mining, this dissertation made enhancements (or improvement) on the algorithm of the positive and negative association rule mining.The main work of the thesis consists of two parts:(1) Combined IUA with SFUA, we conclude a new method that solves the incremental updating problem which is caused when the min- support changed.(2) In the solution of the incremental updating problem, it only needs to compute the frequent itemsets of the updated database, while the negative association rule commonly existed in the infrequent itemsets rather than the frequent itemsets. Therefore, in order to tackle the Incremental Updating Problem of negative association rule mining, it requires to compute both frequent itemsets and infrequent itemsets. The thesis also invent an algorithm which can be used to solve the incremental updating problem of the negative association rule mining with no change of the min-support.