Figure 7
From: Modularity and anti-modularity in networks with arbitrary degree distribution
Comparison of modularity metrics. Comparison of the modularity measures defined in Eqs. (8) and (11). Networks with between 2 and 16 modules were grown depending on the intra-module edge probability π. (For π = 0 the networks are as anti-modular as possible and become k-partite (where k is the number of modules), while for π = 1 they are as modular as possible). (A): Q N [defined in Eq. (8)] for N c = 2 (2 modules, black line), N c = 4 (red), N c = 8 (green), N c = 16 (blue). (B) Q H [defined in Eq. (11)]. Colors as in (A). Each point was averaged over 50 networks with 1,000 nodes. The networks were grown with P N = 0.5, p = 1.0, P E = 1.0, q = 1.0 and P D = 0.0, using undirected edges.