Mean Absolute Error (MAE)

The average loss per example when L1 loss is used. Calculate Mean Absolute Error as follows:

Calculate the L1 loss for a batch.

Divide the L1 loss by the number of examples in the batch.

Click the icon to see the formal math.

For example, consider the calculation of L1 loss on the following batch of five examples:

So, L1 loss is 8 and the number of examples is 5. Therefore, the Mean Absolute Error is:

Mean Absolute Error = L1 loss / Number of Examples

Mean Absolute Error = 8/5 = 1.6

1.
For example, consider the calculation of L1 loss on the following batch of five examples:
2.
So, L1 loss is 8 and the number of examples is 5. Therefore, the Mean Absolute Error is:
3.
Mean Absolute Error = L1 loss / Number of Examples

Created for this library

1.
A retail forecasting team reports MAE per SKU so planners see the typical absolute deviation between forecast and actual sales.
2.
A logistics team reports MAE on delivery time predictions because median absolute error is more intuitive than squared error for operations.
3.
A demand forecasting team uses MAE alongside MAPE so planners can see typical absolute error in units, not just percentages.