Machine Learning
Topic 15 · Model quality · ⏱ about 30 min

✅ Checkpoint 3: Core machine learning

A mixed quiz on the whole core machine learning phase. Pass mark: 6 of 9 on the first try.

Show:
In this topic
Beginner

How this checkpoint works

Answer every question below without looking back at the topics. Only your first try counts. If you reach the pass mark, move on to the next phase. If not, the result tells you which topics to revisit; come back in a few days and try again (the second attempt is good practice, but be honest with yourself about the first).

Check yourself

Quiz

1. Why do we keep a test set we never look at during development?

2. Fitting a scaler on the whole dataset before splitting is…

3. Training error is low and validation error is high. This is…

4. Which regularisation can set weights exactly to zero?

5. Logistic regression outputs…

6. Only 2% of cases are positive. Which metric is LEAST useful on its own?

7. Data is ordered in time. How should you split it?

8. In scikit-learn, which object chains preprocessing and a model so both are fitted on training data only?

9. Your first model should be compared against…

Ready