Mining High Utility Item Sets Through a Swarm-Based Optimization Method

Abstract

Data mining identifies patterns in vast databases to support practical decision-making. Using association rule mining, the patterns in the transaction database are found and reveal client behavior. Frequent Itemset Mining (FIM) Finds Often-Buyd Items. FIM's disdain for item significance is a drawback. Practical applications depend on the relevance of the item. Thus, the High-Utility Itemset Mining (HUIM) challenge requires identifying the most profitable items in the transaction database. High-Utility Items (HUI) in the transaction database can be identified using several methods. HUIM methods based on utility lists are novel and outperform traditional methods in terms of memory consumption and execution time. The main limitation of this algorithm is the costly joins of the utility list. This research presents a highly effective swarm intelligence-based method for optimizing HUIM issues. The execution time of the suggested method is assessed. In addition, it is compared to advanced current methods. Testing on publicly available benchmark data sets shows that the swarm-based strategy outperforms current methods. This technique has wide applications in retail analytics, healthcare, fraud detection, finance, supply chain management, and recommendation systems, enabling businesses and researchers to optimize decision making and extract meaningful insights from transactional data.