inverse_transform()

ScaleTransformer.inverse_transform(states_scaled, inplace=False, locs=None)

Apply the inverse scaling.

Parameters:

states_scaled(n, …) or (p, …) ndarray

Matrix of n-dimensional scaled snapshots, or a single scaled snapshot.

inplacebool

If True, overwrite states_scaled during the inverse transformation. If False, create a copy of the data to untransform.

locsslice or (p,) ndarray of integers or None

If given, assume states_scaled contains the transformed snapshots at only the p indices described by locs.

Returns:

states_unscaled: (n, …) or (p, …) ndarray

Matrix of n-dimensional unscaled snapshots, or the p entries of such at the indices specified by locs.

Raises:

AttributeError

If scaler is a number (not an array) but fit() or fit_transform() have not been called yet.

ValueError

If the states_scaled do not align with the locs (when provided) or the state_dimension (when locs is not provided).