Map equation centrality: community-aware centrality based on the map equation

Abstract To measure node importance, network scientists employ centrality scores that typically take a microscopic or macroscopic perspective, relying on features global structure. However, traditional measures such as degree centrality, betweenness PageRank neglect the community structure found in real-world networks. study importance based flows from mesoscopic we analytically derive community-aware information-theoretic score flow and coding principles behind map equation: equation centrality. Map how much further can compress network’s modular description by not for random walker transitions to respective node, using an adapted scheme determining flow-based point of view. The be determined node’s local context alone because changes only affect other nodes same module. is agnostic chosen model allows researchers select best reflects dynamics process under study. Applied synthetic networks, highlight our approach enables more fine-grained differentiation between than node-local network-global measures. Predicting influential two different dynamical processes networks with measures, find activating tends create largest cascades linear threshold model.