search

nicoloval / NEMtropy

A python3 module for maximum entropy models on networks.
Other
38 stars 5 forks source link

Plans for BiECM? #34

Open deklanw opened 3 years ago

deklanw commented 3 years ago

Any plans to include the bipartite enhanced configuration model?

I know that the DECM can be applied to this case, technically. But, with the BiECM, there are only two variables associated with each node, allowing reduction in unique (degree, strength) pairs.

Thanks!

mat701 commented 3 years ago

We're currently not planning on introducing BiECM, but we could in the future as more models will follow. You're correct: in the case of a bipartite directed (one directional) enhanced configuration model, the reduction is possible. Feel free to fork the project and suggest an implementation!

deklanw commented 3 years ago

Thanks for the response. I went ahead and implemented it in my own library. Honestly, implementing it here seems a bit intimidating looking at the 5k+ line models_functions.py file.

mat701 commented 3 years ago

We didn't implement it since we never actually needed a BiECM, and also weighted models do not actually benefit so much from the reduction, unless you have integer strengths and weights and a power law distribution for both weights and degrees.

By the way, in models functions all functions are independent of each other, so you could add stuff there and create a new weighted/directed bipartite class, but as you prefer!