milia.anomaly.models.mahalanobis#

Robust Mahalanobis distance from a Minimum Covariance Determinant fit.

Classes

Mahalanobis([support_fraction, ...])

Parametric detector: robust Mahalanobis distance to the training distribution.

MahalanobisConfig(*[, name, kind, space, ...])

Validated hyperparameters of the robust Mahalanobis detector.

class milia.anomaly.models.mahalanobis.Mahalanobis(support_fraction=None, random_state=0, contamination=0.1)#

Bases: BaseDetector

Parametric detector: robust Mahalanobis distance to the training distribution.

Parameters:
  • support_fraction (float | None) – Share of points in the MCD support; None uses scikit-learn’s default.

  • random_state (int | None) – Seed for the MCD subset search.

  • contamination (float) – Expected share of outliers, used for the default threshold.

Examples

>>> import numpy as np
>>> rng = np.random.default_rng(0)
>>> train, cal = rng.normal(size=(200, 2)), rng.normal(size=(100, 2))
>>> model = Mahalanobis().fit(train).calibrate(cal, alpha=0.1)
>>> model.flag(np.array([[0.0, 0.0], [8.0, 8.0]])).tolist()
[False, True]
class milia.anomaly.models.mahalanobis.MahalanobisConfig(*, name='mahalanobis', kind='detector', space=SearchSpace(support_fraction=(0.5, 1.0)), support_fraction=None, random_state=0, contamination=0.1)#

Bases: BaseModelConfig

Validated hyperparameters of the robust Mahalanobis detector.

Fields mirror Mahalanobis’s constructor; space is the default Optuna search space. Build the detector with make("detector", "mahalanobis", ...) or Mahalanobis(**config.params()).

Examples

>>> MahalanobisConfig().name
'mahalanobis'