tedana.metrics.dependence.generate_decision_table_score

generate_decision_table_score(*, ascending: List[ndarray] | None = None, descending: List[ndarray] | None = None) → ndarray[source]

Generate a decision table score from an arbitrary set of metrics.

Each metric array is ranked across components. Metrics in descending are ranked so that higher values receive lower (better) scores, while metrics in ascending are ranked so that lower values receive lower (better) scores. The ranks are then averaged per component.

Parameters:

ascending (list of (C,) array_like, optional) – Metric arrays where lower values are better. Ranked with rankdata(x) (rank 1 = smallest value).
descending (list of (C,) array_like, optional) – Metric arrays where higher values are better. Ranked with n - rankdata(x) (rank 0 = largest value).

Returns:

d_table_score ((C) array_like) – Decision table metric scores. Lower values indicate “better” components (e.g., more TE-dependent).