资 源 简 介
Current high utility itemsets mining algorithms adopt two phases: first, generate a large number of candidate itemsets by overestimated utility; second, identify high utility itemsets from the candidates by an additional scan of the original transaction database. The performance bottleneck of these algorithms is the generating & processing of the candidates; and with the increasing of the number of long transaction itemsets and the decreasing of the minimum utility threshold, the situation may become worse. To address this issue, propose a novel tree structure, named TN-Tree (Tail-Node Tree), to maintain the utility information of all transactions on tail-nodes, and an efficient algorithm, named TNT-HUI (Tail-Node Tree based High Utility Itemsets mining), to mine high utility itemsets without generating candidates. The performance of TNT-HUI was evaluated in comparison with the state-of-the-art algorithms on different types of datasets. The experimental results show that TNT-HUI