An efficient join operations for utility list-based high-utility mining approaches using hybrid search technique

Frequent itemset mining (FIM) has firmly established itself as a pivotal and reliable tool in the realm of business analytics, enabling the systematic discovery of valuable patterns and association rules within extensive datasets. However, FIM has a limitation: it only looks at whether an item is present or absent in transactions. This problem is addressed by high-utility itemset mining (HUIM), which considers both the quantity of items and their importance. There are several algorithms for HUIM; their major disadvantage is that they generate too many candidate sets. This problem is tackled by utility-based HUIM algorithms, which are faster and more efficient. However, these algorithms still have an issue with time-consuming join operations. These operations involve unnecessary comparisons when merging two lists of items together. In our study, we introduce a novel technique termed the static increment ratio (SIR) for estimating the probability of an item’s appearance. Additionally, we propose a method known as the hybrid search technique. This technique proficiently compares two item lists through a blend of linear and binary search methods. These approaches leverage the SIR value to make informed decisions, thereby disregarding unnecessary comparisons. We assessed this method using real-world datasets and compared it to state-of-the-art algorithms. The results highlight the method’s reliability and its impressive ability to substantially reduce the execution time of HUIM algorithms, boosting operational efficiency by up to 31%.