Phase 3 · Machine learning · Foundations
Features and labels
Supervised learning: inputs, targets and predictions.
Open this lesson Create a free account
Why it matters
Most practical machine learning in a business is supervised: predict churn, lifetime value, conversion, the right category for a support ticket. Every one of those projects starts by deciding three things: what one example is, which columns are allowed as inputs, and what the target is. Get those wrong (especially by letting the answer leak into the inputs) and the model looks brilliant in testing and fails in real use.
What you'll be able to do
- Name the parts of a supervised problem: examples, features, label, prediction
- Tell regression (predict a number) from classification (predict a category)
- Load a dataset into pandas and split it into X and y
- Spot leakage: features that wouldn't be known when the prediction is made
Before this lesson
Next in this phase: Preparing data