Activity: Create a regularised model comparison note

Estimated active time: 45-75 minutes

Scenario

Use the synthetic learner-support dataset. The fictional platform wants to decide which model families deserve a place in an algorithm portfolio.

Task

Fit ridge and lasso across a range of regularisation strengths and report what each does to coefficients and to validation score.

Required sections

1. Problem and decision.
2. Validation score at three regularisation strengths, for both ridge and lasso.
3. How many coefficients lasso drove to exactly zero, and which.
4. The strength you would choose, and the evidence for it.
5. Keep, delay, or reject against the Module 1 baselines.
6. Limitation note.

Rules

• Standardise features before regularising, or the penalty falls unevenly across columns.
• Do not select the strength on the test set.
• Report which features lasso removed — that list is the interpretable output.

Output

Use the relevant learner-resource template from learner-resources/ or write a one-page memo with the same sections.