Figure 7. Node degree and clustering coefficient distributions.

When the filtering cutoff is set to -∞, all edges returned by the inference network are retained and form a hairball which is characteristic of an exponential network. The connectivity distribution (top) follows a Poisson distribution with λ equal to 150, while the clustering coefficient distribution (bottom) is flat and independent of the node degree. As the cutoff metric is applied to the network, scale-free, hierarchical networks emerge. At a cutoff of 2, the connectivity distribution follows a power law with γ equal to 1.02, and the clustering coefficients begin to scale with the reciprocal of the node degree with an R² value of 0.47. At an even more stringent cutoff value of 2.5, the network further represents a scale-free, hierarchical network where γ equal to 1.19 and R² equal to 0.73. At a cutoff value of 3, the network becomes even more hierarchical with an R² value of 0.96.