Mean absolute error
If we assume the error follows a normal distribution, the log likelihood is . Maximizing this leads to the derivation of MAE.
Analogous to assuming Laplace-distributed errors and Less sensitive to outliers.
MAE
Creator
Creator
Seonglae ChoCreated
Created
2025 Mar 24 15:22Editor
Editor
Seonglae ChoEdited
Edited
2025 Mar 24 15:30Refs
Refs
Mean Squared Error