Рақамли технологияларнинг назарий ва амалий масалалари Том 8 № 2 (2025) · с. 58-64
Graph-based Clustering Algorithm Based on Density Variation
Давронов, Р.Р.
Аннотация
Clustering is one of the main tasks of data analysis aimed at grouping objects into homogeneous subsets without predetermined labels. This article examines the method of column clustering. It uses the concept of iterative removal of low-density nodes to detect “core” nodes (core pixels) and define the structure of clusters. We describe the theoretical foundations of the method, provide implementation details, and analyze the obtained results on synthetic datasets (including those created using the scikit-learn library). Furthermore, we compare the proposed algorithm with other known clustering methods using the ARI (Adjusted Rand Index) metric. Experiments show that this approach effectively identifies structures of different shapes and densities and demonstrates competitive results compared to classical methods.
кластеризация на графахлокальная плотностьудаление узлов графаназначение кластероввариация плотностиclustering in graphslocal densityremoving graph nodespurpose of clustersdensity variation
Источник метаданных: OAI-PMH архив журнала · Sindex не хранит полный текст, а даёт ссылку на источник.