scienxlab / redflag

Safety net for machine learning pipelines. Plays nice with sklearn and pandas.
https://scienxlab.org/redflag
Apache License 2.0
22 stars 6 forks source link

Flag classes not seen in training, or not present in evaluation #84

Open kwinkunks opened 1 year ago

kwinkunks commented 1 year ago

Flag if see [A, B, C] in training then [A, B] or [A, B, C, D] in evaluation.

Would go in target.py I think, and have sklearn component.

kwinkunks commented 1 year ago

I think sklearn flags these, but maybe the error is sometimes obscure? Check first.

noobCoderVP commented 1 year ago

I would like to work on this issue, I have good experience of working with SKlearn for machine learning, can you elaborate a little about the issue? is it like the same rows should not be present in training and testing? Assign me this issue if yes