The inherent community structure of hyperbolic networks

A remarkable approach for grasping the relevant statistical features of real networks with help random graphs is offered by hyperbolic models, centred around idea placing nodes in a low-dimensional space, and connecting node pairs probability depending on distance. It widely appreciated that these models can generate are small-world, highly clustered scale-free at same time; thus, reproducing most fundamental common networks. In present work, we focus less well-known property popularity-similarity optimisation (PSO) model $\mathbb{S}^1/\mathbb{H}^2$ from this family, namely generated approaches also contain communities wide range parameters, which was certainly not an intention design models. We extracted studied using well-established community finding methods such as Louvain, Infomap label propagation. The observed high modularity values indicate structure become very pronounced under certain conditions. addition, modules found different algorithms show good consistency, implying indeed apparent structural units. Since appearance rather representing systems well, feature makes them even more suitable describing than thought before.