probly.conformal_prediction.scores.aps.common

APS Score implementation with optional Randomization (U-term).

Functions

register(cls, func)

Register a implementation for a specific type.

Classes

APSScore(model[, randomize, random_state])

Adaptive Prediction Sets (APS) nonconformity score.

class probly.conformal_prediction.scores.aps.common.APSScore(model, randomize=True, random_state=None)

Bases: ClassificationScore

Adaptive Prediction Sets (APS) nonconformity score.

Parameters:

• model (Predictor)

• randomize (bool)

• random_state (int | None)

calibration_nonconformity(x_cal, y_cal, probs=None)

Compute calibration scores (vectorized, backend-agnostic).

Parameters:

• x_cal (Sequence[Any])

• y_cal (Sequence[Any])

• probs (Any | None)

Return type:

npt.NDArray[np.floating]

compute_score(probs)

Calculate APS scores with randomization U-term.

Parameters:

probs (Any)

Return type:

Any

predict_nonconformity(x_test, probs=None)

Compute scores for all labels (stays on original device).

Parameters:

• x_test (Sequence[Any])

• probs (Any | None)

Return type:

Any

probly.conformal_prediction.scores.aps.common.register(cls, func)

Register a implementation for a specific type.

Parameters:

• cls (LazyType)

• func (Callable)

Return type:

None