drop1#

SuperGLM.drop1(
X: object,
y: NDArray,
sample_weight: NDArray | None = None,
offset: NDArray | None = None,
*,
test: str = 'Chisq',
) → DataFrame#

Drop-one deviance analysis for each feature.

At exactly zero fitted dispersion, a zero deviance change has statistic zero and p-value one; a nonzero deviance change is undefined and raises.