This paper introduces a memory-efficient approach to parallel community detection in large graphs, utilizing weighted Misra-Gries sketches to reduce the memory footprint of Louvain, Leiden, and Label Propagation Algorithms (LPA) at the cost of moderate runtime increase.
Edge covariates improve community detection in the VEC-SBM model.
The authors propose improvements to community detection algorithms by incorporating random walks, enhancing efficiency and maintaining complexity. Their approach aims to refine clustering results while validating the effectiveness through experiments.
The authors present two overlapping network community detection algorithms based on extended modularity and cosine functions, applicable to both undirected and directed graphs.