Precision vs recall decisions, imbalance, leakage, scaling, validation mistakes, bias vs variance diagnosis and the name traps that cost marks in ML courses and interviews.
Precision or recall: which matters more for (a) cancer screening and (b) a spam filter that deletes mail?
Tap to flip
Back
(a) Recall: missing a cancer (false negative) is the worse mistake. (b) Precision: wrongly deleting a real email (false positive) is the worse mistake.