A classifier can make two different positive-class mistakes. A false positive raises an alert for a negative case. A false negative misses a positive case. Accuracy counts correct predictions across all rows, which can conceal a weak positive-class result when negatives are common.
Precision asks how many alerts were correct. Recall asks how many actual positive cases were found. F1 combines precision and recall through their harmonic mean. The right metric depends on what the team needs to control, and none of them tells you whether a probability estimate is well calibrated. Use class 1 as the positive outcome in these tasks. When a denominator is zero, the requested zero_division policy makes the result explicit rather than allowing an undefined value to derail the analysis.