Efficient mining top-k high utility itemsets in incremental databases based on threshold raising strategies and pre-large concept

High utility itemset mining (HUIM) is a sub-problems of frequent itemset mining (FIM) that has received a lot of interest from researchers. It is used to analyze user behavior and improve business efficiency. The top-k high utility itemsets mining (top-k HUIM) issue aims to explore the k-itemsets with the highest utility from the database to handle the difficulty of threshold selection. Top-k HUIM algorithms ignore the transactions continuously added to the database in a dynamic environment, resulting in inaccurate top-k HUI results. However, the current top-k HUIM algorithms in the incremental database require users to request mining manually, or else, have it automatically processed every time the incremental batch is scanned, which is very small compared to the original database. Re-mining when the data is not updated enough affects the results and consumes a lot of resources without obtain new valuable insights. This research presents a raising threshold strategy to take advantage of the original database's mining results combining the updated database strategies. Furthermore, the paper proposes definitions of top-k mining using pre-large concept, thresholds, conditions for re-mining and method to solve the problem of always mining. Combining the proposed techniques and strategies, a complete "PreTK" algorithm is proposed to solve the proposed issues. The experiments are deployed to compare the algorithm's performance on diverse databases with baseline algorithms. The results demonstrate that the proposed method outperforms the state-of-the-art algorithms and may provide results faster, even when remining is necessary.