Regression
A supervised task that predicts a continuous numeric value rather than a category.
- Outputs a real number: price, revenue, temperature, time.
- Quality measured by error (MAE, RMSE), not accuracy.
- Linear regression is the canonical algorithm.
When to use: When the target is a "how much" question: forecasts, pricing, risk scores, ETA.
Example: Predict the sale price of a house given square footage, bedrooms, neighborhood, and year built. Output is dollars; quality is RMSE on held-out homes.