Mining High-Utility Itemsets Based on Multiple Minimum Support and Multiple Minimum Utility Thresholds

Mining high utility itemsets from a transactional database refer to the discovery of high utility itemsets that generate high profit and several approaches have been proposed for this task in recent years. Algorithms like HUIM-MMU and MHU-Growth overcome the limitation of using a single threshold for the whole database. However, they still generate a large number of candidate itemsets and thus it degrades the performance of the algorithms. In this paper, we address this issue by combining two different kinds of thresholds used by HUIM-MMU and MHU-Growth. By using these two thresholds we propose two algorithms namely HUIM-MMSU and HUIM-IMMSU. HUIM-MMSU is a candidate generation and retest based algorithm, which relies on sorted downward closure (SDC) property. On the other hand, HUIM-IMMSU uses a tree-like data structure. Experiment result shows that the proposed two algorithms can effectively discover high utility itemsets from the transactional database.