Disentangling group and link persistence in Dynamic Stochastic Block models
Inference from single snapshots of temporal networks can misleadingly group communities if the links between snapshots are correlated.
P. Barucca, F. Lillo, P. Mazzarisi, D. Tantari
We study the inference of a model of temporal networks in which both communities and links keep memory of previous network state. By considering maximum likelihood inference from single snapshot observation of the network, we show that link persistence decreases the detectability threshold, preventing the inference of communities even when they are in principle strong enough, while community persistence tends to restore this possibility. Then we show that the inferred communities share a maximum overlap with those of a specific previous instant of time, corresponding to the maximum of a time-lagged assortativity parameter, and therefore they can be closer to those of the past than of the present. We analytically characterize these transitions as a function of the memory parameters of the model.
More in Hidden communities
Naturally occurring networks have an underlying scale-free structure that is often clouded by finite-size effects in the sample data.
A new tool derived from information theory quantitatively identifies trees, hierarchies and community structures within complex networks.
An iterative version of a method to identify hierarchies and rankings of nodes in directed networks can partly overcome its resolution limit.
The spectral density of graph ensembles provides an exact solution to the graph partitioning problem and helps detect community structure.