1 paper
Juri Opitz, Sebastian Burst
The 'macro F1' metric is frequently used to evaluate binary, multi-class and multi-label classification problems. Yet, we find that there exist two different formulas to calculate…