On 09.11.2016 15:37, lucidyan wrote:
Hi! I'm trying to perform community detection algorithm with /nested stochastic block model/ and I had some questions:
1. Is there a way to take into account the weights of the edges? I read in *BlockState* about some /weights/, but did not understand what it is and how to use them.
These are positive integer-valued property maps that should contain the edge multiplicities between nodes.
2. How to get the hierarchical results from the /nested stochastic block model/ to obtain a kind of agglomerative dendrogram?
Look at NestedBlockState.draw() and draw_hierarchy().
3. How to obtain numerical values of each vertex belonging to the clusters, after margins collectioning, like it drew in pie vertex chart in examples?
They are stored in the property maps returned by collect_vertex_marginals().
4. How to measure quality of partitioning: is it any other metrics, than was in examples? Is it necessary that both/Model Evidence/ was bigger and /Bethe Entropy/ was lower? What if only one metrics performs better, then other degrades?
The Bethe approximation is more accurate, and it should be trusted more in these corner cases. If you want to compare individual partitions, only the description length is necessary. Best, Tiago -- Tiago de Paula Peixoto <tiago@skewed.de>