SklearnModelAdapter.score#

SklearnModelAdapter.score(X, y, *, coords=None, sample_weight=None, multioutput='raw_values', force_finite=True)[source]#

Return per-output \(R^2\) scores from the sklearn model.

Parameters:

  • X (Any) – Predictor matrix.

  • y (Any) – Observed outcomes.

  • coords (dict[str, Any] | None) – Ignored for sklearn backends.

  • sample_weight (Any | None) – Sample weights passed to sklearn.metrics.r2_score().

  • multioutput (Literal['raw_values']) – The required aggregation mode. Per-unit scores require the raw value for each output.

  • force_finite (bool) – Whether to replace non-finite scores for constant targets, passed to sklearn.metrics.r2_score().

Returns:

One unit_{i}_r2 entry per output. Point estimates carry no dispersion entries.

Return type:

pd.Series