Authors: Rashmi Dutta Baruah; Plamen Angelov
Addresses: School of Computing and Communications, Lancaster University, Lancaster, LA1 4WA, UK ' School of Computing and Communications, Lancaster University, Lancaster, LA1 4WA, UK
Abstract: In this paper, we attempt to detect the changes in the structure of an evolving social network. We define a novel measure to quantify the dynamics of the network and use it to generate a timeline for the network that indicates the changes. We consider that any significant change in the network is as a result of occurrence of some event, and thus, identify the day or time step when the event took place. Further, the analysis involves identification of key individuals and their close associates that were active on that day. Finally, the case study involves identification of individuals having communication behaviour similar to the key individuals. To group such individuals, we use a recently proposed online clustering approach called evolving clustering (eClustering). We demonstrate the efficacy of the proposed quantification measures by conducting several experiments and comparing the results with the ground truth.
Keywords: evolving social networks; dynamic social networks; social network analysis; SNA; online clustering; evolving clustering; cell phones; mobile phones; network dynamics; communication behaviour; key individuals.
International Journal of Social Network Mining, 2013 Vol.1 No.3/4, pp.254 - 279
Available online: 03 Feb 2014 *Full-text access for editors Access for subscribers Purchase this article Comment on this article