Can be the estimated Mean Absolute Error (MAE) equal or more than 1? If it is possible (which it happened to me), when it happens and what is the reason?


  • $\begingroup$ The question might be backwards. If all your errors are expected to have absolute value less than 1, you can conclude that the MAE will be less than 1 also. $\endgroup$ – Eduardo Elael Nov 21 '18 at 14:54

Of course an absolute error can be greater than $1$. The "absolute" says you are calculating a difference.

When it happened to you it probably means just that the data are very widely spread out.

If you are still puzzled by anomalies in your data you should probably post them in a new question, show us your analysis and what bothers you.

  • $\begingroup$ Bokler. Thank you for the answer. I got the idea $\endgroup$ – user3233712 Nov 21 '18 at 15:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.