A Hybrid GA–PSO–CSA with Bit-Vector Optimization for High-Utility Itemset Mining
| dc.contributor.author | Almeida, Tracy | |
| dc.date.accessioned | 2026-08-07T06:28:41Z | |
| dc.date.available | 2026-08-07T06:28:41Z | |
| dc.date.issued | 2026 | |
| dc.description.abstract | High Utility Itemset Mining (HUIM) identifies itemsets based on utility, such as profit or significance, making it valuable in retail, healthcare, and finance. Conventional methods like Two-Phase and UP-Growth struggle with exponential search space, repeated database scans, and high memory use, limiting performance on large datasets. This paper introduces a hybrid metaheuristic combining Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Crow Search Algorithm (CSA), enhanced with Bit-Vector Optimization (BVO) for efficient itemset representation. GA enables exploration, PSO accelerates convergence, CSA maintains diversity, while BVO reduces computation cost. Tests on IBM synthetic datasets and the Retail-FIMI benchmark show the proposed GA–PSO–CSA with BVO outperforms traditional and standalone metaheuristics in convergence speed, execution time, and solution quality, proving its scalability and effectiveness for real-world HUIM tasks. | |
| dc.identifier.uri | http://rcca.ndl.gov.in/handle/123456789/636 | |
| dc.language.iso | en | |
| dc.publisher | Data Science and Security | |
| dc.title | A Hybrid GA–PSO–CSA with Bit-Vector Optimization for High-Utility Itemset Mining | |
| dc.type | Book chapter |