Modularity clustering is force-directed layout

Two natural and widely used representations for the community structure of networks are clusterings, which partition the vertex set into disjoint subsets, and layouts, which assign the vertices to positions in a metric space. This paper unifies prominent characterizations of layout quality and clustering quality, by showing that energy models of pairwise attraction and repulsion subsume Newman and Girvan’s modularity measure. Layouts with optimal energy are relaxations of, and are thus consistent with, clusterings with optimal modularity, which is of practical relevance because both representations are complementary and often used together.

💡 Research Summary

The paper “Modularity clustering is force‑directed layout” establishes a rigorous connection between two widely used representations of community structure in networks: discrete clusterings that partition vertices into disjoint groups, and continuous layouts that embed vertices in a metric space. The authors begin by formalizing weighted graphs, where each vertex v carries a non‑negative weight w_v and each unordered pair {u,v} has an edge weight w_{uv}. They then introduce the (a, r)‑energy model for layouts, a family of force‑based quality measures parameterized by two real constants a (attraction exponent) and r (repulsion exponent) with a > r. The energy of a layout p is defined as

E_{a,r}(p) = Σ_{u≠v}