PageRank centrality for temporal networks (Q822935)

This paper studies PageRank centrality for temporal networks with discrete-time scale. A temporal network is proposed as a multilayer network and a fourth-order tensor is used to obtain the representation of the multilayer network with \(N\) nodes over \(T\) time layers. The centrality is based upon the tensor equations across all these layers, for which an iterative algorithm is adopted. By calculating the eigenvector of a multi-homogeneous map, the centrality values of nodes and time layers are obtained. Numerical simulations have been performed to test the proposed algorithm. Three temporal networks are considered, including the Email-Eu-core temporal network and college messaging temporal network.