آمار

تعداد نشریات11
تعداد شماره‌ها228
تعداد مقالات2,317
تعداد مشاهده مقاله3,623,847
تعداد دریافت فایل اصل مقاله2,663,839
Eyvazi, Hossein, Badzohreh, Mohammad, Shahrokhi, Seyed Ali. (1404). VelvetFlow: An engineering pipeline for robust multi-density clustering. فناوری آموزش, 10(4), 333-358. doi: 10.22061/jdma.2025.12039.1131
Hossein Eyvazi; Mohammad Badzohreh; Seyed Ali Shahrokhi. "VelvetFlow: An engineering pipeline for robust multi-density clustering". فناوری آموزش, 10, 4, 1404, 333-358. doi: 10.22061/jdma.2025.12039.1131
Eyvazi, Hossein, Badzohreh, Mohammad, Shahrokhi, Seyed Ali. (1404). 'VelvetFlow: An engineering pipeline for robust multi-density clustering', فناوری آموزش, 10(4), pp. 333-358. doi: 10.22061/jdma.2025.12039.1131
Eyvazi, Hossein, Badzohreh, Mohammad, Shahrokhi, Seyed Ali. VelvetFlow: An engineering pipeline for robust multi-density clustering. فناوری آموزش, 1404; 10(4): 333-358. doi: 10.22061/jdma.2025.12039.1131

VelvetFlow: An engineering pipeline for robust multi-density clustering

Journal of Discrete Mathematics and Its Applications
مقاله 3، دوره 10، شماره 4، اسفند 2025، صفحه 333-358    اصل مقاله (3.42 M)
نوع مقاله: Full Length Article
شناسه دیجیتال (DOI): 10.22061/jdma.2025.12039.1131
نویسندگان
Hossein Eyvazi* ؛ Mohammad Badzohreh؛ Seyed Ali Shahrokhi
Department of Computer Science, University of Tarbiat Modares, Tehran, I. R. Iran
تاریخ دریافت: 20 اردیبهشت 1404،  تاریخ بازنگری: 15 تیر 1404،  تاریخ پذیرش: 18 مرداد 1404 
چکیده
Problem. Real-world datasets seldom respect a single density scale: tight blobs, elongated ribbons, and isolated points often coexist. Classical algorithms such as DBSCAN or \textit{k}-means require domain-specific parameter tuning and provide only ad-hoc support for anomaly detection.
Solution. We introduce VelvetFlow, an engineering pipeline that turns a set of well-understood building blocks into a cohesive, end-to-end workflow for multi-density clustering \emph{and} principled outlier detection. The pipeline is composed of three reusable stages:
(i) \emph{Contextual-density splitting} assigns every point to a high- or low-density partition using a single neighbourhood size $k$.
(ii) \emph{Density-aware clustering} applies a Jaccard-guided \textit{FusedNeighbor}+DBSCAN routine to the sparse partition and HDBSCAN to the dense partition-without introducing new hyper-parameters.
(iii) \emph{Scaled-MST verification} re-examines the complete $k$-NN graph, flags weakly connected components, and validates them with a $k$-NN gate; this step recovers small remote clusters while filtering genuine anomalies.
کلیدواژه‌ها
multi-density clustering؛ outlier detection؛ HDBSCAN؛ DBSCAN؛ MST؛ fused neighbor
آمار
تعداد مشاهده مقاله: 81
تعداد دریافت فایل اصل مقاله: 62


سامانه مدیریت نشریات علمی. طراحی و پیاده سازی از سیناوب