ITHUIM: An efficient algorithm for mining inter-transaction high utility itemsets

Data mining is an active and vital research field that has led to the development of numerous algorithms and solutions. One significant contribution of this field is the discovery of high utility itemsets (HUIs) in transactional databases, which provides managers with valuable insights for maximizing profit. However, existing work on HUI mining has focused exclusively on the relationship among items within individual transactions, while ignoring relationships that occur across consecutive transactions. To address this limitation, we extend high utility itemset mining by considering mega-transaction relationships. We introduce the concept of mega-transaction and provide related definitions, formally defining the problem of mining inter-transaction high utility itemsets (ITHUIs). Furthermore, we develop and prove several mathematical properties that form the theoretical foundation for practical ITHUI mining. Based on these properties, a candidate pruning technique exploiting extended items is proposed to reduce the search space. Finally, we design a novel mining algorithm for efficiently discovering ITHUIs, named ITHUIM (ITHUI Miner).