Skip to content
BENIBIX Learn

Mean Absolute Error (MAE)

AI📏 Evaluation & BenchmarksRegression Metrics

Also known as MAE

The average size of a model's misses when it guesses numbers. Lower is better.

Think of it like

It is like guessing how many sweets are in some jars. You add up how far off each guess was.

Example

Jack's model guesses each day's top temperature in London. Over 30 days, its guesses are off by 1.5 degrees on average, so its MAE is 1.5.

Why it matters

It is the easiest error score to explain. Anyone can read it as the usual miss, in real units.

Formula

MAE = average of |true value - guess|

Definition

Mean absolute error, or MAE, measures how far a model's guesses are from the true numbers. For each guess, you take the gap and drop any minus sign. Then you take the average of all those gaps. The result is in the same unit as what you predict, such as euros or degrees. An MAE of 2 degrees means the guesses are off by 2 degrees on average. Lower is better, and 0 means every guess was exact.
MAE is the average length of the gaps. Each gap counts by its own size.

Common mistake

Comparing MAE across tasks with different units. An MAE of 5 is tiny for house prices and huge for shoe sizes.

Key takeaways

  1. 1
    Average gap. The size of each miss, sign dropped.
  2. 2
    Real units. It reads in euros, degrees or days.
  3. 3
    Every miss counts. Each one by its own size.

Compare