Module 1 Assessment: Algorithm Selection Readiness

Estimated active time: 30-45 minutes

Question 1

What is the majority-class baseline on this dataset, and why is it the first number to record?

Pass answer: gives the figure and explains that any model below it is worse than a constant prediction.

Question 2

Why must the split be fixed now rather than per module?

Pass answer: recognises that results across modules are only comparable on an identical split.

Question 3

What margin over the logistic-regression baseline should an algorithm clear to earn a place?

Pass answer: names a specific margin and ties it to the cost of added complexity.

Question 4

Why is no tuning performed in this module?

Pass answer: identifies that a reference point tuned against itself is not a reference point.