Katz centrality, control analysis and directed networks
Kieran Sharkey (University of Liverpool)
Thursday 15th February, 2018 14:00-15:00 311B Mathematics and Statistics Building
Due to train cancellations this talk is no longer going ahead.
Network centrality measures are used to identify the most important nodes that govern dynamics across networks. Effective identification of these nodes has applicability in, for example, social network structures, epidemic contact networks and biochemical networks. In this talk I particularly focus on Katz centrality, first proposed by Leo Katz in 1953. Katz centrality is directly related to eigenvector centrality and to Google’s PageRank algorithm which I shall also discuss.
I will present a different interpretation of Katz centrality as a steady-state solution to a continuous-time dynamical system which I refer to as Katz dynamics. By applying a control analysis to Katz dynamics, a more detailed understanding of centrality on directed networks is obtained.