I performed a rolling window (i.e. do full sample, then next 4 observations until the last, and so on...) k-fold test for out of sample testing due to limited number of observations.

From the MAPE results, it seems like that with sample size it increases and sort of reaches a ceiling at a certain sample size.

For the MAE results, the MAE falls until the middle of the sample and starts rising again the larger the sample gets.

I'm new to this and am not sure how to interpret these results and would appreciate any help on this issue.


MAE and MAPE results (X axis - sample size)

  • $\begingroup$ Pls post a graph of your data. $\endgroup$ – horaceT Sep 13 '16 at 20:39
  • $\begingroup$ No, not graphs of the MAE/MAPE, but graph of one of your time series. Here is the reason. MAPE is a percentage, where it has numerator and a denominator. A constant error of your model may show up as a higher MAPE if the base level of your time series drops over time. Since you didn't tell us the nature of these time series, it's hard for anyone to help you. $\endgroup$ – horaceT Sep 14 '16 at 16:43
  • $\begingroup$ Ah, I misunderstood. Just uploaded the Time series $\endgroup$ – rayven1lk Sep 15 '16 at 18:13

Your Answer

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

Browse other questions tagged or ask your own question.