Sociomeeting: Dynamical Metrics for Temporal Networks: State of art, Extensions and Applications

Temporal networks can be viewed as time series of networks. This interpretation raises the question if and how existing tools for the analysis of time series can be extended to temporal networks. In this talk, I will briefly summarize work already done in this area and present new results on the extension of tools from dynamical system theory for unlabelled graphs. The extension of these tools, such as the autocorrelation function and Lyapunov exponents, to unlabelled graphs is particularly relevant in cases where labels on nodes are not available, either due to privacy reasons or matching problems encountered while building the network. I will also discuss the analysis of real social temporal networks.