1. High-utility itemsets mining integrating an improved crow search algorithm and particle search optimization.
- Author
-
Ledmi, Makhlouf, Ledmi, Abdeldjalil, Souidi, Mohammed El Habib, Hamdi-Cherif, Aboubekeur, Maarouk, Toufik Messaoud, and Hamdi-Cherif, Chafia Kara-Mohamed
- Subjects
- *
METAHEURISTIC algorithms , *SEARCH algorithms , *MATHEMATICAL optimization , *DATA mining , *ALGORITHMS - Abstract
High-utility itemsets mining (HUIM) is an important research topic in data mining that consists of extracting efficient itemsets that generate high profit in databases. Metaheuristic algorithms are good candidates for finding some sufficiently good high-utility itemsets (HUIs) while reducing the prohibitive computational cost incurred by such mining. However, metaheuristic algorithms suffer from crucial issues: runtime and memory consumption, entailing scalability issues. To contribute in alleviating these drawbacks, we explore a hybrid approach combining metaheuristics with an optimization technique, and experimentally show the obtainment of better quality solutions. The exploration component in our approach relies on the crow search algorithm (CSA) which is a method that presents a strong capacity in the global search process. The exploitation component is based on the particle swarm optimization (PSO) which is an established stochastic optimization technique known to have a good convergence capability. This article proposes the HUIM-ICSO algorithm, coupling CSA and PSO, as a hybrid method for mining HUIs. In addition to the proposed main CSA-based algorithm, two additional novel enhanced variants are incorporated and integrated with PSO. Experiments are undertaken on 12 publicly available datasets and compared with state-of-the-art algorithms, such as mHUIMiner, HUIM-SPSO and HUIM-AF. Experiments show that, as the number of transactions or items grows, the suggested approach remains scalable and delivers favorable outcomes in terms of runtime, memory usage, and the quantity of HUIs mined, all while maintaining a good convergence rate. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF