Residuals
Difference between observed and predicted Y — what the model does not explain.
Also known as: residual, prediction error
A residual is the leftover the regression cannot explain: the distance from the observed Y to the prediction .
- Residual plot (e_i vs. ) — should show a random cloud around 0
- Pattern / curvature → non-linear effects missing in the model
- Funnel shape → heteroscedasticity, standard errors unreliable
- Q-Q plot of residuals — check for normality
See also
Used in
In the Algorithm Lab
Sources
- Montgomery, D. C., Peck, E. A. & Vining, G. G. — Introduction to Linear Regression Analysis, 6th Edition, Chapter 4
- Cook, R. D. & Weisberg, S. — Residuals and Influence in Regression, Chapman & Hall