Danica Greetham: Abstract

In this talk, we will  look at the problem of centrality ranking in dynamic networks. Centrality is an intuitive concept that tells us how important or involved is an individual node in a network. Many different definitions exist but not all will work in evolving networks where edges are created and destroyed in each time step. We are focusing here on one that works well, slightly modified, in evolving networks, Katz centrality, and show how it can be calculated efficiently for relatively large networks.

We then use this calculation to inform a decision on a network based intervention and report the results from several small scale human studies where we observed (and interfered with) the mood dynamics of evolving communication networks.

Organised by BCS SGAI The Specialist Group on Artificial Intelligence http://www.bcs-sgai.org