EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

A Taxonomy of DBSCAN Variants Addressing Parameter Dependence and High-Dimensionality

Anunaya Manoj, Masood Husain Siddiqui, Deepak Kumar Singh

Interdisciplinary Journal of AI, Machine Learning & Data Science · 03.06.2026 · 1 (2) · e013

DOI: 10.66261/s5y5tk64 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Interdisciplinary Journal of AI, Machine Learning & Data Science, online erschienen am 03.06.2026. Verfasst von Anunaya Manoj und 2 weiteren Autor:innen. Die Arbeit nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Density-based clustering algorithms have emerged as a significant area of research within unsupervised machine learning due to their capability to identify clusters based on data density. Among these methods, the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm has gained considerable attention owing to its robustness against noise and outliers. However, DBSCAN’s sensitivity to input parameters, its inability to detect clusters of varying densities, and its poor performance in high-dimensional spaces limit its applicability. Considering these challenges, this review categorizes variants of DBSCAN into four major groups: (1) parameter-independent, (2) parameter-reduced, (3) high-dimensional, and (4) hybrid algorithms. Within each category, we provide a detailed description of parameter estimation and clustering strategies. Our analysis highlights the use of graph-based methodologies, tree-based and grid structures, and other techniques including RNN, LSH, DLT, and triangular inequality to help automate parameter estimation and accelerate neighborhood searches. Moreover, hybrid strategies incorporating feature selection, feature extraction, graph-based modeling, game theory, and deep learning have shown promising potential in addressing both challenges simultaneously. By consolidating insights across these research directions, this review offers a comprehensive perspective on the evolution of DBSCAN, equipping researchers and practitioners with a deeper understanding of its limitations, existing solutions, and avenues for future exploration.

Zitieren

APA

Manoj, A., Siddiqui, M. H., & Singh, D. K. (2026). A Taxonomy of DBSCAN Variants Addressing Parameter Dependence and High-Dimensionality. Interdisciplinary Journal of AI, Machine Learning & Data Science, 1(2), e013. https://doi.org/10.66261/s5y5tk64

Vancouver

Manoj A, Siddiqui MH, Singh DK. A Taxonomy of DBSCAN Variants Addressing Parameter Dependence and High-Dimensionality. Interdisciplinary Journal of AI, Machine Learning & Data Science. 2026;1(2):e013. doi:10.66261/s5y5tk64

BibTeX
@article{Manoj2026,
  author = {Manoj, Anunaya and Siddiqui, Masood Husain and Singh, Deepak Kumar},
  title = {A Taxonomy of DBSCAN Variants Addressing Parameter Dependence and High-Dimensionality},
  journal = {Interdisciplinary Journal of AI, Machine Learning & Data Science},
  year = {2026},
  volume = {1},
  number = {2},
  pages = {e013},
  doi = {10.66261/s5y5tk64}
}
Kostenlos starten →