pybispectra.utils.compute_rank

pybispectra.utils.compute_rank(data: ndarray, sv_tol: int | float = 1e-05) → int

Compute the minimum rank of data from non-zero singular values.

Parameters:

datanumpy.ndarray, shape of [epochs, channels, times]

Data to find the rank of.

sv_tolint | float (default 1e-5)

Tolerance to use to define non-zero singular values, based on the largest singular value. Singular values greater than the largest singular value multiplied by the tolerance are considered to be non-zero.

Returns:

rankint

Minimum rank of data over epochs.