3 ms·
For the first two, these relationships can be modelled from the derived community graph of a binary network, so I am interested to see the performance gains ove
by dleeftink 3y ago
For the first two, these relationships can be modelled from the derived community graph of a binary network, so I am interested to see the performance gains over the binary approach.
See for instance, the ngraph coarsen procedure that reduces communities and their interrelations to single nodes connected to other community nodes [0].
[0]: https://github.com/anvaka/ngraph.coarsen https://github.com/anvaka/ngraph.coarsen
- zmgsabst 3y agoI too would be curious to see the comparison of different approaches. But my intuition is that caching derived graphs and caching computed cells will be similar — and might even be two conceptual frameworks on equivalent data.
- woolion 3y agoNot only that but there are lots of good results and algorithms on 'bipartite graphs', which are the kind that encode hypergraphs, so in practice it should work very well.