Question: After examining the model’s mistakes, you decide to add new input variables. What should those variables be chosen to do?

Sample answer: They should be built to cut down the specific pattern of mistakes that the review uncovered, rather than to improve all errors equally.

Key points:

• Use the review to identify a recurring mistake pattern.
• Add features that reduce that particular kind of error.

Rubric: Must state that the added variables are meant to reduce a specific error pattern or class of mistakes found in the review.