check_2d#

SuperLSS.check_2d(
X: object | EagerFrame,
y: NDArray,
covariate: str | NDArray,
other: str | NDArray,
*,
names: tuple[str, str] | None = None,
sample_weight: NDArray | None = None,
offsets: Mapping[str, NDArray] | None = None,
seed: int = 42,
**kwargs: Any,
) → BinnedCheck2D#

Return the mean residual on a two-dimensional grid of two covariates.

Each covariate is a column name of X or an array with one value per row. The tile means say where in the joint region the fit drifts, which a pair of one-dimensional checks cannot show; n_bins is a pair, one count per axis.