BPt.extensions.SingleConnectivityMeasure.inverse_transform#

SingleConnectivityMeasure.inverse_transform(connectivities, diagonal=None)[source]#

Returns connectivity matrices from connectivities, vectorized or not.

If kind is ‘tangent’, the covariance matrices are reconstructed.

Parameters
connectivitieslist of n_subjects numpy.ndarray with shapes (n_features, n_features) or (n_features * (n_features + 1) / 2,)

or ((n_features - 1) * n_features / 2,) Connectivities of each subject, vectorized or not.

diagonalnumpy.ndarray, shape (n_subjects, n_features), optional

The diagonals of the connectivity matrices.

Returns
outputnumpy.ndarray, shape (n_subjects, n_features, n_features)

The corresponding connectivity matrices. If kind is ‘correlation’/ ‘partial correlation’, the correlation/partial correlation matrices are returned. If kind is ‘tangent’, the covariance matrices are reconstructed.