Macro-Recall

Classification metric computing the per-head macro-averaged recall: recall is taken per class and then averaged with equal weight per class. Because every class counts the same regardless of how many samples it has, this diverges sharply from plain accuracy on an imbalanced evaluation set, where a dominant class can carry accuracy on its own.

Classes with no support in the evaluation set are left out of the average rather than counted as zero. This is the same quantity as balanced accuracy in its unadjusted form.

Configuration

Selected as name: macro-recall. Takes no options.

services:
  metrics:
    select:
      - name: macro-recall
        kwargs: {}