Skip to content

Different r2_score multioutput default in r2_score and base.RegressorMixin #12772

Closed
Closed
Different r2_score multioutput default in r2_score and base.RegressorMixin#12772
@qinhanmin2014

Description

@qinhanmin2014
qinhanmin2014
opened on Dec 14, 2018

We've changed multioutput default in r2_score to "uniform_average" in 0.19, but in base.RegressorMixin, we still use multioutput='variance_weighted' (#5143).
Also see the strange things below:

scikit-learn/sklearn/multioutput.py

Lines 283 to 286 in 4603e48

# XXX remove in 0.19 when r2_score default for multioutput changes
from .metrics import r2_score
return r2_score(y, self.predict(X), sample_weight=sample_weight,
multioutput='uniform_average')

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions