CHUIM: An Efficient Algorithm for Correlated High Utility Itemset Mining

Mining High-Utility Itemsets (HUIs) stands as a pivotal task in database analysis, like customer transactions, to identify itemsets of high importance. Despite the development of several traditional algorithms over the years for this purpose, they often produce an excessive number of High Utility Itemsets (HUIs). Unfortunately, many of these HUIs lack meaningful correlations between their items, making them unreliable for decision-making as their high utility might be due to random chance. To tackle this issue, our paper introduces the Correlated High Utility Itemset Miner (CHUIM). CHUIM utilizes the concept of productivity in both normal and exclusive domains to eliminate HUIs that occur randomly without genuine item correlations. Experimental evaluation on benchmark databases demonstrates that CHUIM efficiently prunes HUIs lacking inherent item correlations, thereby enhancing the reliability of the generated results.