CovarianceEstimator

class probly.layers.numpy.CovarianceEstimator(*args, **kwargs)

Bases: Protocol

An estimator exposing a precision matrix after being fitted.

Structurally satisfied by the estimators in sklearn.covariance, e.g. EmpiricalCovariance, LedoitWolf or ShrunkCovariance.

fit(features: ndarray, /) → object

Fit the estimator on a feature matrix of shape (N, feature_dim).

precision_: ndarray