Study Of A Kind Of Mining Association Rule Finding And Its Updating Algorithm Based Searching Algorithm

Mining association rules and their updates are major aspect of data mining Domain. Recently, there have many algorithms proposed for efficient discovery of association rules in large database, which all come into their own. There are different incremental updating algorithms based on different mining association rules. First, this paper introduces a mining association rule algorithm based on searching algorithm, which can be called NewQAIS. By comparing to Apriori algorithm, we can realize that NewQAIS algorithm is better in some aspects. This paper also analyses the composing character of item sets, points out some drawbacks of NewQAIS algorithm. The drawbacks of NewQAIS algorithm lie in: on the one side , the efficiency of it creating item sets and getting the support is weak; on the other hand, although the concept of the aggregated set is proposed in this algorithm, yet there is no a practical method pointed out in this paper. On the basis of the comprehension about the QAIS algorithm, a method to solve the above drawbacks is pointed out, which leads to the NewQAIS algorithm. Compared to QAIS algorithm, the NewQAIS algorithm is of its own character. The algorithm considers not only mining association rules but also updating mining association rules. The updating algorithm based on NewQAIS algorithm refuses the old way which differentiates frequent sets from item sets by singly computing the support of the item sets, and basing on the data structure of the aggregated itemset, it classifies the item sets to differentiate them as a whole, so the updating algorithm is of the availability and superiority.