Section checkpoint

Evaluate & Ship

9 questions drawn from the 3 chapters in this section. If most land on the first try, you are ready for the next section. If not, the explanation after each answer says which idea to revisit.

Checkpoint

9 questions · first-try answers earn the most XP

1. Why is cross-validation more trustworthy than a single train/validation split?

2. What does early stopping do to a model during the training loop?

3. A regularisation hyperparameter like max_depth or C controls what?

4. What does the elbow plot help you decide when choosing K for K-Means?

5. Why does the chapter say a cluster nobody can act on is decoration?

6. What is the main practical use of clustering when you already have labelled data?

7. Your training accuracy is 0.99 and validation accuracy is 0.62. What does this most likely indicate?

8. Which set should you keep untouched until you have finished choosing a model?

9. You deploy a model and predictions look obviously wrong. Your service loads the pipeline from disk. What is the first thing to check?

Checkpoint

What this checkpoint covers