Machine Learning · easy

Explain precision vs recall with an example.

Asked in Data Scientist interviews, in the Machine Learning round.

Short answer

Precision = of predicted positives how many correct; recall = of actual positives how many caught.

How to answer it

Precision: of the things you flagged, how many were right. Recall: of the things that were really there, how many you flagged.

The example to carry: a spam filter that quarantines 100 emails, of which 90 are spam. Precision is 90%. If the inbox actually received 300 spam emails that day, recall is 90 out of 300, 30%. The filter is careful and misses a lot.

Raise recall by flagging more aggressively and precision falls, because more of what you flag is legitimate mail. The two move against each other through the threshold, and which one you protect is a product decision:

• Spam filter: protect precision. A lost real email is worse than a spam that gets through.
• Cancer screening: protect recall. A missed case is worse than a follow-up test.
• Fraud blocking: depends on the cost of a blocked customer versus a fraudulent charge, so compute it.

from sklearn.metrics import precision_score, recall_score
precision_score(y_true, y_pred), recall_score(y_true, y_pred)

F1 is their harmonic mean, useful when you must report one number and have no cost model; F-beta lets you weight recall more (beta 2) or less (beta 0.5). Say plainly that both are computed at a threshold, and that the PR curve shows every threshold at once.

• What they are checking: correct definitions, and an example where you say which to protect and why.
• Common mistake: swapping them. Anchor on "precision is about what you predicted; recall is about what was true".

Related questions

• What metrics would you use for an imbalanced fraud model?
• How do you handle class imbalance in classification?
• Explain L1 vs L2 regularization.
• Explain how a random forest works.
• How would you detect data leakage in a pipeline?
• What is gradient boosting and how does it differ from bagging?

Practice this for real

• Check your resume against a Data Scientist job description — free, no account needed.
• Every data interview question, by role and topic
• Scored AI mock interviews — two free, no card