When the class distribution is highly unbalanced for a binary classification task, which of the following statements is NOT true (choose one) when evaluating a classifier: Accuracy is a better metric than F1-score. Accuracy will be high when all examples are predicted as belonging to the majority class. Recall for the majority class will be 1.0 when all examples are predicted as belonging to the majority class. F1-score should be used as metrics in addition accuracy