pitcp.utils.hpd_volume

hpd_volume(hpd, X, confidence_level=0.9, n_samples=1000)[source]

Estimates conditional HPD-set volumes by Monte Carlo integration.

Parameters:

  • hpd (HPD) – Calibrated HPD estimator.

  • X (np.typing.ArrayLike) – Features with shape (n_samples, n_features).

  • confidence_level (float | Sequence[float], optional) – Coverage level or levels. Defaults to 0.9.

  • n_samples (int, optional) – Monte Carlo samples per feature row. Defaults to 1000.

Returns:

Estimated volumes with shape (n_samples,) or ``(n_samples,

n_levels)``.

Return type:

np.ndarray