Using a two-phase evolutionary framework to select multiple network spreaders based on community structure

Yu Hsiang Fu, Chung Yuan Huang*, Chuen-Tsai Sun

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

1 Scopus citations

Abstract

Using network community structures to identify multiple influential spreaders is an appropriate method for analyzing the dissemination of information, ideas and infectious diseases. For example, data on spreaders selected from groups of customers who make similar purchases may be used to advertise products and to optimize limited resource allocation. Other examples include community detection approaches aimed at identifying structures and groups in social or complex networks. However, determining the number of communities in a network remains a challenge. In this paper we describe our proposal for a two-phase evolutionary framework (TPEF) for determining community numbers and maximizing community modularity. Lancichinetti–Fortunato–Radicchi benchmark networks were used to test our proposed method and to analyze execution time, community structure quality, convergence, and the network spreading effect. Results indicate that our proposed TPEF generates satisfactory levels of community quality and convergence. They also suggest a need for an index, mechanism or sampling technique to determine whether a community detection approach should be used for selecting multiple network spreaders.

Original languageEnglish
Pages (from-to)840-853
Number of pages14
JournalPhysica A: Statistical Mechanics and its Applications
Volume461
DOIs
StatePublished - 1 Nov 2016

Keywords

  • Community detection
  • Genetic algorithm
  • Multiple network spreaders
  • Network spreading
  • Social network analysis

Fingerprint Dive into the research topics of 'Using a two-phase evolutionary framework to select multiple network spreaders based on community structure'. Together they form a unique fingerprint.

Cite this