A K-Means, Ward, and DBSCAN Repeatability Study
Cloud Computing and Data Science · 02.04.2026 · 215-45
DOI: 10.37256/ccds.7220269455 ↗Kurz gefasst
Fachartikel in der Zeitschrift Cloud Computing and Data Science, online erschienen am 02.04.2026. Verfasst von Anthony Bertrand und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Reproducibility is essential in machine learning because it ensures that a model or experiment yields the same scientific conclusion. For specific algorithms, repeatability with bitwise identical results is also a key for scientific integrity because it allows debugging. We decomposed several very popular clustering algorithms: K-Means, Density-Based Spatial Clustering of Applications with Noise (DBSCAN), and Ward into their fundamental steps, and we identify the conditions required to achieve repeatability at each stage. We use an implementation example with the Python library scikit-learn to examine the repeatable aspects of each method. Our results reveal non-repeatable behavior with K-Means when the number of OpenMP threads exceeds two. This work aims to raise awareness of this issue among both users and developers, encouraging further investigation and potential fixes.
Zitieren
APA
Bertrand, A., Nguifo, E. M., Antoine, V., & Hill, D. R. (2026). A K-Means, Ward, and DBSCAN Repeatability Study. Cloud Computing and Data Science, 215-45. https://doi.org/10.37256/ccds.7220269455
Vancouver
Bertrand A, Nguifo EM, Antoine V, Hill DR. A K-Means, Ward, and DBSCAN Repeatability Study. Cloud Computing and Data Science. 2026:215-45. doi:10.37256/ccds.7220269455
BibTeX
@article{Bertrand2026,
author = {Bertrand, Anthony and Nguifo, Engelbert Mephu and Antoine, Violaine and Hill, David R.C.},
title = {A K-Means, Ward, and DBSCAN Repeatability Study},
journal = {Cloud Computing and Data Science},
year = {2026},
pages = {215--45},
doi = {10.37256/ccds.7220269455}
}
EUVIMED