3 Jul
2020
3 Jul
'20
6:23 p.m.
Am 03.07.20 um 20:17 schrieb Davide Cittaro:
Would it make sense to get a graph from partitions (where edges are weighted on connectivity among them) and apply NSBM then? This would produce a mixed model where the deepest level is actually a PP, and the hierarchy doesn't assume any constraint on assortativity.
I'm not sure this makes sense. The nested SBM is a generative model where the upper layers are the parameters of the layer below. The PP model does not take a full matrix of connections between groups as parameters, it has only the number of edges inside communities, and outside. So those two models do not fit together. -- Tiago de Paula Peixoto <tiago@skewed.de>