Training loss decreases but validation loss increases. What's happening and the fix?

Overfitting; add regularization / early stopping

Underfitting; use a bigger model

Your positive class is 0.5% of data. How do you build a good model?

Use PR-AUC + class weighting + threshold tuning

How do you choose the decision threshold for a binary classifier in practice?

Optimize on validation for business cost

Quiz · Foundations & Concepts

Applied ML Case Studies quiz

3 multiple-choice questions on Applied ML Case Studies — every answer comes with an explanation, so each question teaches something even when you miss it.

Professional3 questions

Ready to test yourself on Applied ML Case Studies?

3 quiz questions available. Pick a quiz length — questions are shuffled every run.

Review the lesson first

Test yourself