CategoricalCrossValidation

CategoricalCrossValidation(
    actual: ndarray,
    probabilities: ndarray,
    names: list[str],
)

Cross-validation of categorical indicator kriging: actual holds the category codes, probabilities the (samples, categories) corrected probabilities and names the categories; NaN where a sample had too few neighbors.

brier property

brier

Mean squared difference between each category's probability and indicator; 0 is perfect.

most_likely property

most_likely

Code of the most probable category, ties to the lowest; NaN where unestimated.