Regression

Regression is a supervised learning task that predicts a continuous numerical value rather than a category. Examples include estimating a home’s price, forecasting energy demand, predicting delivery time, or measuring how a treatment may change an outcome. Linear regression models a straight-line relationship, while trees, neural networks, and other methods can capture more complex patterns. Evaluation may use mean absolute error, mean squared error, or domain-specific measures that reflect the cost of mistakes. Predictions should include uncertainty where possible, especially beyond the range of training data. Correlation learned by regression does not establish causation, and strong benchmark performance does not protect against changing real-world conditions.