Can we automatically find hidden communities in tangled networks?

Researchers combined edge betweenness (a measure of how often connections pass through a node) with dendrogram analysis to automatically identify nested community structures in networks. Unlike supervised methods that need human labels, this approach works blind—building a hierarchy from the network's natural flow patterns. They tested it on fractal networks (which have communities within communities) and partition networks, showing it can both detect clusters and measure how modular they are.