Residual
noun
say it: “rih-ZIJ-oo-ul” /rɪˈzɪdʒuəl/
The difference between a real value and the value a model predicted for it.
What is a residual?
A residual is one example's miss: the actual value minus the model's guess. It's positive when the guess was too low and negative when it was too high. Error scores such as mean absolute error are built from the residuals.
cups = [40, 65, 80]
guessed = [45, 60, 82]
print([c - g for c, g in zip(cups, guessed)]) # [-5, 5, -2]
In a sentence
- “Plotting each residual showed the model guessed too low on the hottest days.”
- “A residual of -5 means the model guessed 5 cups too many.”
Level: Journeyman · More AI Basics terms at this level
Updated