Computer scienceData scienceMachine learningRegressionRegression metrics

MAE

Over- and underestimations

Report a typo

Can we tell if the model overestimates or underestimates the true target values by the MAE scores?

Select one option from the list

We can, because the metric is not sensitive to outliers

We can, because all errors proportionally contribute towards the score

We can't, there is no continious derivative

We can't, the usage of absolute values makes it impossible to tell the difference

___

Create a free account to access the full topic