Discovering natural communities in networks

摘要: Abstract Understanding and detecting natural communities in networks have been a fundamental challenge networks, science generally. Recently, we proposed hypothesis that homophyly/kinship is the principle of based on real network experiments, model to explore selection nature evolving, measure structure entropy networks. Here community finding algorithm by our We found exactly identifies almost all generated selection, if any, or precisely approximates planted existing models. verified very well ground-truth some world are semantically well-defined, naturally finds balanced communities, may larger modularity than algorithms modularity, for Our provides first time an approach analyzing true results demonstrate minimization large-scale