Community Structure

Community structure is the partition of a network into subsets of nodes — communities or modules — that are densely connected internally and sparsely connected to other subsets. It is the organizational principle by which real-world networks achieve functional differentiation: in biological networks, communities correspond to functional pathways; in social networks, to groups and institutions; in information networks, to topic clusters. The detection of community structure is one of the central algorithmic problems in network science, and it remains unresolved because the very definition of a community depends on the null model against which density is measured.

The most widely used approach is modularity optimization, which compares the density of intra-community edges to what would be expected in a random network with the same degree distribution. But this method has a resolution limit — it systematically fails to identify communities smaller than a scale set by the total network size. This means that the method finds communities, but not all communities, and the ones it misses may be precisely the communities that matter for understanding local dynamics.

The deeper question is whether community structure is a property of the network or a property of the algorithm used to find it. If different algorithms return different partitions of the same network, then community structure is not a topological invariant but an interpretive frame. This would not make it useless — interpretation is not the enemy of science — but it would make community detection a choice among models rather than a discovery of structure.