A community merger of optimization algorithm to extract overlapping communities in networks

Qi Li, Jiang Zhong, Qing Li, Chen Wang, Zehong Cao

Research output: Contribution to journalArticlepeer-review

8 Scopus citations

Abstract

A community in networks is a subset of vertices primarily connecting internal components, yet less connecting to the external vertices. The existing algorithms aim to extract communities of the topological features in networks. However, the edges of practical complex networks involving a weight that represents the tightness degree of connection and robustness, which leads a significant influence on the accuracy of community detection. In our study, we propose an overlapping community detection method based on the seed expansion strategy applying to both the unweighted and the weighted networks, called OCSE. First, it redefines the edge weight and the vertex weight depending on the influence of the network topology and the original edge weight, and then selects the seed vertices and updates the edges weight. Comparisons between OCSE approach and existing community detection methods on synthetic and real-world networks, the results of the experiment show that our proposed approach has the significantly better performance in terms of the accuracy.

Original languageEnglish
Article number8555542
Pages (from-to)3994-4005
Number of pages12
JournalIEEE Access
Volume7
DOIs
StatePublished - 2019
Externally publishedYes

Keywords

  • complex network
  • dense subgraph
  • Overlapping community detection
  • weighted network

Fingerprint

Dive into the research topics of 'A community merger of optimization algorithm to extract overlapping communities in networks'. Together they form a unique fingerprint.

Cite this