Evolution in Social Networks: A Survey
There is much research on social network analysis but only recently did scholars turn their attention to the volatility of social networks. An abundance of questions emerged. How does a social network evolve – can we find laws and derive models that expla
- PDF / 613,954 Bytes
- 27 Pages / 439.37 x 666.142 pts Page_size
- 108 Downloads / 192 Views
Abstract
There is much research on social network analysis but only recently did scholars turn their attention to the volatility of social networks. An abundance of questions emerged. How does a social network evolve – can we ¿nd laws and derive models that explain its evolution? How do communities emerge in a social network and how do they expand or shrink? What is a community in an evolving network – can we claim that two communities seen at two distinct timepoints are the same one, even if they have next to no members in common? Research advances have different perspectives: some scholars focus on how evolution manifests itself in a social network, while others investigate how individual communities evolve as new members join and old ones become inactive. There are methods for discovering communities and capturing their changes in time, and methods that consider a community as a smoothly evolving constellation and thus build and adapt models upon that premise. This survey organizes advances on evolution in social networks into a common framework and gives an overview of these different perspectives.
Keywords:
Dynamic social networks, social network evolution, community evolution, stream clustering, incremental tensor-based clustering, dynamic probabilistic models
1.
Introduction
Evolution in social networks is a research domain of some intricacy. Understanding the evolution of social communities is an appealing subject, but the evolution of the social networks themselves is a distinct, no less captivating problem. The underpinnings of social network evolution are to be found in modeling and studying evolving graphs. In contrast, an evolving community is not necessarily part of an evolving graph: many scholars rather perceive a community as a group of individuals that have some features in common. These C. C. Aggarwal (ed.), Social Network Data Analytics, DOI 10.1007/978-1-4419-8462-3_6, © Springer Science+Business Media, LLC 2011
150
SOCIAL NETWORK DATA ANALYTICS
different perceptions have lead to different methods for the study of community evolution and to different de¿nitions of the concept "evolving community". This study is a survey of the advances on understanding and on predicting the evolution of social constellations, sometimes called social networks, other times called communities and often modeled as graphs. Informally, evolution refers to a change that manifests itself across the time axis. In the ¿eld of Knowledge Discovery from Data, there is a distinction between mining static data and mining a data stream. The stream paradigm of computation dictates that instances arrive in sequential order and each instance is seen only once [22]. Stream mining algorithms must solve the challenge of adapting a model over the new data under resource constraints. Some stream mining algorithms focus on monitoring model evolution, i.e. understanding how the model changes. Stream mining algorithms that adapt or monitor evolving communities are obviously within the scope of our survey, albeit graph stream mi
Data Loading...
