What metrics would you use to evaluate the performance of a classification model?
Instruction: Discuss at least three different metrics and when they are most appropriately used.
Context: This question assesses the candidate's understanding of model evaluation and their ability to select appropriate metrics based on the problem context.
Updated
Official answer available
Read the opening below, then unlock the full answer and practical guidance.
Choose metrics from the intended decision, class prevalence and error costs, then evaluate on representative held-out data.
Accuracy is the fraction of correct predictions. It is a useful overall view when the class mix and mistake costs make that summary meaningful, but can hide poor performance on a rare positive class. Precision is the fraction of positive predictions that are truly positive...