AAI Handbook
← All models
algorithm

Linear Regression

beginnersupervised
parametricregressiontabular

Linear Regression predicts a number by adding weighted feature values and a baseline offset.

Mechanisms

weighted sumleast squares

Properties

extrapolates linearlyinterpretable

Constraints

sensitive to outliers

Practical profile

Explainability
high
Training cost
low
Inference cost
low
Data appetite
low
Concepts to understand
GeneralizationLoss and OptimizationRegularizationTrain, Validation, and Test Sets