0

I am trying to use minimize_blockmodel_dl to identify communities in a weighted (undirected) graph.

I found out that this method once supported a eweight parameter, but it doesn't anymore.

Is there any method in graph-tool nowadays which takes into account weights and automatically looks for the optimal number of components?

Pietro Battiston
  • 7,930
  • 3
  • 42
  • 45

1 Answers1

2

This is described in detail in the documentation: https://graph-tool.skewed.de/static/doc/demos/inference/inference.html#edge-weights-and-covariates

(Also, the eweight parameter is still there, it is only accessed differently. However, this parameter always meant edge multiplicities, not arbitrary weights. The link above describes how to incorporate arbitrary weights, which can be real, signed, etc.)

Tiago Peixoto
  • 5,149
  • 2
  • 28
  • 28