A statistical model where a response is estimated as a linear function of predictors.
Linear operations include:
- Adding vectors (+)
- Multiplying vectors by a constant (*)
- \(y = mx + b\) (grade school)
- \(Y_i = \beta_0 + \beta_1 X_i + \epsilon_i\) (grown-up)
An observation (i) has a mean value that is estimated by an intercept coefficient (\(\beta_0\)) and a slope coefficient (\(\beta_1\)) multiplied by a predictor (\(X\)) with some random error (\(\epsilon_i\))