Log Loss as Evaluation Metric

Hard

machine-learning Airbnb Netflix Lyft

Two classifiers both achieve 90% accuracy on a binary task. Classifier A has a log loss of 0.25, while Classifier B has a log loss of 0.55. What does this difference reveal?

Multiple Choice

Classifier B is better because higher log loss indicates more confident predictions Classifier A produces better-calibrated probability estimates, assigning higher confidence to correct predictions Log loss is unreliable when both models have the same accuracy The difference is meaningless since accuracy is the true measure of performance

Correct!

Incorrect — try again next time!

Upgrade to Pro to unlock solutions and explanations

Log Loss as Evaluation Metric

Multiple Choice

Related problems

Deep-dive guides