Cargando…

Isolate sets partition benefits community detection of parallel Louvain method

Community detection is a vital task in many fields, such as social networks, and financial analysis, to name a few. The Louvain method, the main workhorse of community detection, is a popular heuristic method based on modularity. But it is difficult for the sequential Louvain method to deal with lar...

Descripción completa

Detalles Bibliográficos
Autores principales: Qie, Hang, Li, Shijie, Dou, Yong, Xu, Jinwei, Xiong, Yunsheng, Gao, Zikai
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9114431/
https://www.ncbi.nlm.nih.gov/pubmed/35581228
http://dx.doi.org/10.1038/s41598-022-11987-y
Descripción
Sumario:Community detection is a vital task in many fields, such as social networks, and financial analysis, to name a few. The Louvain method, the main workhorse of community detection, is a popular heuristic method based on modularity. But it is difficult for the sequential Louvain method to deal with large-scale graphs. In order to overcome the drawback, researchers have proposed several parallel Louvain methods (Parallel Louvain Method, PLM), which suffer two challenges: (1) latency in the information synchronization and (2) communities swap. To tackle these two challenges, we propose a graph partition algorithm for the parallel Louvain method. Different from existing graph partition algorithms, our graph partition algorithm divides the graph into subgraphs called isolate sets, in which vertices are relatively decoupled from others, and the PLM computes and synchronizes information without delay and communities swap. We first describe concepts and properties of isolate sets. In the second place, we propose an algorithm to divide the graph into isolate sets, which enjoys the same computation complexity as the breadth-first search. Finally, we propose the isolate-set-based parallel Louvain method, which calculates and updates vertices information without latency and communities swap. We implement our method with OpenMP on an 8-cores PC. Experiments on 18 graphs show that our parallel method achieves a maximum 4.62 [Formula: see text] speedup compared with the sequential method, and outputs higher modularity on 14 graphs.