Better internal structure for `QDForest`

python-qds / qdscreen

Quasi-determinism screening for fast Bayesian Network Structure Learning (from T.Rahier's PhD thesis, 2018)

https://python-qds.github.io/qdscreen/

BSD 3-Clause "New" or "Revised" License

2 stars 0 forks source link

Better internal structure for `QDForest` #12

Open smarie opened 3 years ago

smarie commented 3 years ago

Currently QDForest can either be created from an adjmat or a parents array. Is this the most optimal ? We should probably use a scipy.sparse structure or two instead.

This should be tailored to usage:

get_arcs method should be efficient for QDSelectorModel.fit(X)
walk_arcs method should be efficient for QDSelectorModel.predict_qd(X)

smarie commented 3 years ago

Also concerning names (and dtypes in the future ?), it is probably better to do all the work on arrays, and to remember a map index -> name and dtype.

smarie commented 3 years ago

https://github.com/caesar0301/treelib

and

https://networkx.org/

might help also