High Utility

High-utility pattern mining (HUPM) aims to identify subsets of data exhibiting high value or profit, going beyond simple frequency analysis. Current research focuses on improving the efficiency of HUPM algorithms, particularly for sequential data and large datasets, employing techniques like optimized data structures (e.g., sequence projections, utility lists), tighter upper bounds for pruning the search space, and novel pruning strategies. These advancements enhance the scalability and practicality of HUPM, enabling its application in diverse fields such as predictive modeling (e.g., ICU admission prediction) and recommendation systems, where identifying high-value patterns is crucial for decision-making.