Figure 1. A schematic visualization of our approach to study properties of network populations.

First, we map networks to feature vectors. Then we analysis these feature vectors with a hierarchical clustering. The resulting clustering allow us to conclude back to the similarity of the network populations.