Open sami6mz opened 1 year ago
Besides, could it be a good idea to shuffle the folds?
kf = KFold(n_splits=crossfit,shuffle=True)
Anyway data dimulation is already randomized.
No need then.
@houssamzenati I think this is included in the refactor you plan to do, right?
In benchmark_mediation.py, there are 4 chunks of the following code :
Which I think we could replace by something more straightforward :
Or even :
wdyt?
Note : First version of the code returns a list of lists, while the second returns a list of tuples.