Identifying influential users by their postings in social networks

Beiming Sun, Vincent To Yee Ng

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

20 Citations (Scopus)


Much research effort has been conducted to analyze information from social networks, including finding the influential users. In this paper, we propose a graph model to represent the relationships between online posts of one topic, in order to identify the influential users. Besides the role of starters, we suggest a new role, the connecter, to help bridging two different clusters of posts. Three methods for measuring the influences of online posts are discussed to distinguish starters and connecters in the graph. The results of the different measurements can then be integrated to determine the most influential posts and their respective authors. With the information of the explicit and implicit relationship between posts, our model tries to identify the most influential users based on their direct interaction as well as the implicit relationship among postings. The experiment is performed on Twitter to verify the model and the three methods of influence measurement. The interpretation of the methods is also given to justify the experiment results.
Original languageEnglish
Title of host publicationMSM'12 - Proceedings of 3rd International Workshop on Modeling Social Media
Number of pages8
Publication statusPublished - 25 Jul 2012
Event3rd International Workshop on Modeling Social Media, MSM'12 - Milwaukee, WI, United States
Duration: 25 Jun 201225 Jun 2012


Conference3rd International Workshop on Modeling Social Media, MSM'12
Country/TerritoryUnited States
CityMilwaukee, WI


  • Graph entropy
  • Graph modeling
  • Node centrality
  • Social networks
  • Starter and connecter identification
  • User influence

ASJC Scopus subject areas

  • Modelling and Simulation


Dive into the research topics of 'Identifying influential users by their postings in social networks'. Together they form a unique fingerprint.

Cite this