Not every prediction is a category. Sometimes the answer is a number: tomorrow's temperature, a house price, a patient's blood pressure. Regression predicts continuous values from input features.

Linear regression fits a straight line or hyperplane. Polynomial regression adds curvature. Tree-based methods like random forests and gradient boosting handle nonlinear relationships without explicit feature engineering.

Common regression metrics

• Mean absolute error
• Mean squared error
• Root mean squared error
• R-squared, proportion of variance explained

Outliers distort squared-error metrics more than absolute-error ones. Choosing the right loss depends on whether large errors deserve extra weight or not. Domain knowledge usually settles the question.