Phương pháp khai thác tập mục hữu ích cao từ luồng dữ liệu dựa trên thuật toán bầy voi

Abstract

— High Utility Itemset Mining (HUIM) is a prominent research direction in the field of data mining. Traditional HUIM algorithms often struggle with the exponential explosion of the search space, leading to limitations in scalability. To address this issue, heuristic-based HUIM algorithms have attracted significant attention. However, these methods are prone to premature convergence, resulting in the omission of potentially valuable itemsets. To overcome these limitations, we propose a new algorithm named SHUIM_HE, based on the Elephant Herding Optimization (EHO) algorithm, to efficiently mine high-utility itemsets from data streams in resource-constrained environments. The core innovation of the algorithm lies in a position evolution strategy based on female elephant, which significantly reduces the search space and improves the algorithm's execution performance. Experiments conducted on real-world datasets demonstrate that the proposed algorithm outperforms state-of-the-art heuristic HUIM algorithms.