An approach in the direction of extraction of the high utility patterns from certain data by utilizing tree

High Utility Pattern Mining (HUPM) has advanced beyond traditional frequent pattern mining by capturing the unique profits associated with items, thus revealing more valuable patterns. Despite this advancement, conventional HUPM approaches assume perfect database integrity, which can undermine the reliability of mined results in the presence of unknown errors or noise. This paper introduces a novel HUPM model designed to address these limitations. The proposed model, named Positive and Negative unit utilities-based High Utility Patterns from Certain Data (PNHUPC), integrates a High Utility Pattern Tree (HUPT) to enhance mining effectiveness and reduce the necessity for multiple database scans. The HUPT structure efficiently retains transactional information and high utility itemsets, thereby optimizing the mining process. Initially, the model converts an uncertain dataset into a certain dataset by incorporating factors such as positive and negative unit profit, purchase quantity, seasonal frequency, and item probability. This conversion results in a refined list of items with attributes reflecting their potential utility. Subsequently, unpromising candidates are strategically pruned from this list. The construction of the HUPT further improves mining performance, making the proposed approach a robust solution for extracting valuable patterns in the presence of data imperfections.