Model of community emergence in weighted social networks

J. M. Kumpula*, J. P. Onnela, J. Saramäki, J. Kertész, K. Kaski

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract (may include machine translation)

Over the years network theory has proven to be rapidly expanding methodology to investigate various complex systems and it has turned out to give quite unparalleled insight to their structure, function, and response through data analysis, modeling, and simulation. For social systems in particular the network approach has empirically revealed a modular structure due to interplay between the network topology and link weights between network nodes or individuals. This inspired us to develop a simple network model that could catch some salient features of mesoscopic community and macroscopic topology formation during network evolution. Our model is based on two fundamental mechanisms of network sociology for individuals to find new friends, namely cyclic closure and focal closure, which are mimicked by local search-link-reinforcement and random global attachment mechanisms, respectively. In addition we included to the model a node deletion mechanism by removing all its links simultaneously, which corresponds for an individual to depart from the network. Here we describe in detail the implementation of our model algorithm, which was found to be computationally efficient and produce many empirically observed features of large-scale social networks. Thus this model opens a new perspective for studying such collective social phenomena as spreading, structure formation, and evolutionary processes.

Original languageEnglish
Pages (from-to)517-522
Number of pages6
JournalComputer Physics Communications
Volume180
Issue number4
DOIs
StatePublished - Apr 2009
Externally publishedYes

Keywords

  • Communities
  • Social network
  • Weighted network

Fingerprint

Dive into the research topics of 'Model of community emergence in weighted social networks'. Together they form a unique fingerprint.

Cite this