DistanceBasedCredalSet

class probly.representation.credal_set.DistanceBasedCredalSet(*args, **kwargs)[source]

Bases: CategoricalCredalSet, Generic

A credal set defined by a distance metric around a central distribution.

abstract property barycenter: T

Return the barycenter of the credal set.

abstractmethod classmethod from_sample(sample: Sample[T]) → Self[source]

Create a credal set from a finite sample.

Parameters:

sample – The sample to create the credal set from.

Returns:

The created credal set.

abstract property nominal: T

The nominal distribution.

abstract property num_classes: int

Return the number of classes in the credal set.

protected_axes: ClassVar[dict[str, int]] = {'nominal': 0, 'radius': 0}
abstract property radius: Any

The distance from the nominal distribution.

type: ClassVar[Literal['categorical']] = 'categorical'

Examples using probly.representation.credal_set.DistanceBasedCredalSet

Conformalized Credal Set Prediction — PyTorch.

Conformalized Credal Set Prediction — PyTorch.

Plotting credal sets on the simplex

Plotting credal sets on the simplex

Plotting credal sets on a spider (radar) chart

Plotting credal sets on a spider (radar) chart

Distance-based credal set

Distance-based credal set