mne_denoise.dss.selection.auto_select_components_robust

mne_denoise.dss.selection.auto_select_components_robust(eigenvalues: ndarray, sigma: float = 3.0, knee_rel_floor: float = 0.01, knee_min_ratio: float = 3.0) → int

Combine outlier and knee component counts and return the larger count.

Parameters:

eigenvaluesarray-like

DSS component scores in descending order.

sigmafloat, default=3.0

Outlier threshold multiplier.

knee_rel_floorfloat, default=0.01

Relative knee floor.

knee_min_ratiofloat, default=3.0

Minimum knee ratio.

Returns:

int

Larger of the two proposed counts.