An Optimization Driven Consensus Framework for Ensemble Clustering Using Particle Swarm Optimization
Management Strategies and Engineering Sciences · 2026 · 8 (5) · 1-13
DOI: 10.61838/msesj.357 ↗Kurz gefasst
Fachartikel in der Zeitschrift Management Strategies and Engineering Sciences, 2026. Verfasst von Abdolrashid Rezvani und 4 weiteren Autor:innen.
Abstract (Originalsprache)
Clustering ensemble methods aim to improve robustness by integrating multiple base partitions; however, many existing consensus strategies rely on heuristic aggregation and remain sensitive to instability in individual clustering results. To address this limitation, this paper proposes Consensus Particle Swarm Clustering (CPSC), an unsupervised ensemble clustering framework that formulates consensus construction as an explicit optimization problem in the label space. In the proposed approach, multiple base clustering solutions are transformed into a unified label based representation, and Particle Swarm Optimization is employed to search for a consensus partition that maximizes agreement among ensemble members independently of the original feature space. This design enhances robustness against initialization sensitivity and variability across clustering algorithms. The effectiveness of CPSC is evaluated on standard benchmark datasets, including Iris, Diabetes, Yeast, Two Spiral, and Ralf rings. Experimental results demonstrate that CPSC consistently outperforms individual clustering methods and conventional consensus techniques in terms of on the. In particular, the proposed method achieves error rates as low as 4.00% on the Iris dataset and 31.25% on the Diabetes dataset, while yielding an average error rate of 32.32% across all evaluated benchmarks. Sensitivity analysis further indicates that CPSC exhibits stable performance with respect to the number of clusters, converging within an effective range of K. Although the framework requires the number of clusters to be specified in advance and introduces additional computational cost due to iterative optimization, the results confirm the effectiveness of optimization driven consensus formation. Overall, CPSC provides a robust and extensible solution for unsupervised ensemble clustering, offering a principled alternative to heuristic consensus methods.
Zitieren
APA
Rezvani, A., Mirzaei, A., Mikaeilvand, N., Nouri-Moghaddam, B., & Gudakahriz, S. J. (2026). An Optimization Driven Consensus Framework for Ensemble Clustering Using Particle Swarm Optimization. Management Strategies and Engineering Sciences, 8(5), 1-13. https://doi.org/10.61838/msesj.357
Vancouver
Rezvani A, Mirzaei A, Mikaeilvand N, Nouri-Moghaddam B, Gudakahriz SJ. An Optimization Driven Consensus Framework for Ensemble Clustering Using Particle Swarm Optimization. Management Strategies and Engineering Sciences. 2026;8(5):1-13. doi:10.61838/msesj.357
BibTeX
@article{Rezvani2026,
author = {Rezvani, Abdolrashid and Mirzaei, Abbas and Mikaeilvand, Nasser and Nouri-Moghaddam, Babak and Gudakahriz, Sajjad Jahanbakhsh},
title = {An Optimization Driven Consensus Framework for Ensemble Clustering Using Particle Swarm Optimization},
journal = {Management Strategies and Engineering Sciences},
year = {2026},
volume = {8},
number = {5},
pages = {1--13},
doi = {10.61838/msesj.357}
}
EUVIMED