Chapter 1: Classical ML
Framing: what to predict, for whom, when, and what to beat
A model can score well and still reward the wrong behaviour. A clear frame helps you defend each choice in interviews.
What you will learn
- Translate a product goal into a label, prediction unit and prediction time
- Choose a dummy baseline and state the bar a model must clear
- Spot a proxy label that rewards the wrong behaviour
Key terms
- label
- the outcome a prediction task targets
- unit
- the subject of one prediction
- horizon
- the future window a prediction covers
- baseline
- a simple prediction method whose score provides a comparison
Prerequisites
Take Welcome to Classical ML: a model from start to finish first. It comes right before this lesson in the course.
Where this lesson picks up: You chose what to predict. What makes that choice precise?
What comes next
How does a model learn your chosen target from rows?
Next lesson: Fitting a line: what 'best fit' means and how to read the slope
More lessons in this part
- Welcome to Classical ML: a model from start to finish 7 min
- Fitting a line: what 'best fit' means and how to read the slope 20 min
- The exact best line: two equations, and MAE 20 min
- From a line to a probability: the sigmoid 7 min
- Training logistic regression: log loss, its gradient, and why not squared error 7 min
- Thresholds, precision and recall: the cut-off is a product decision 16 min
- Honest splits and leakage: random, group and time-based 18 min
Take the lesson
Short cards, one idea at a time, with quick checks as you go.