In this paper, we propose an algorithm to handle the transaction insertion for efficiently updating the discovered high average-utility upper-bound itemsets (HAUUBIs) based on the average-utility (AU)-list structure a...
详细信息
ISBN:
(纸本)9789811064876;9789811064869
In this paper, we propose an algorithm to handle the transaction insertion for efficiently updating the discovered high average-utility upper-bound itemsets (HAUUBIs) based on the average-utility (AU)-list structure and the Fast UPdated (FUP) concept. The proposed algorithm divides the HAUUBIs existing in the original database and new transactions into four cases, and each case can be respectively maintained to identify the actual high average-utility itemsets (HAUIs) without multiple database scans and enormous candidate generation. Experiments showed that the proposed algorithm has better performance compared to state-of-the-art algorithm in terms of runtime and generates the similar number of candidates.
暂无评论