I have monthly time series and I need to make predictions for the next quarter. However, for operational reasons, the forecast must be made one month before the quarter. That way, I need, in practice, to make a forecast of four months, sum up the four forecasts and subtract the stock of the month before the beginning of the quarter, yielding the forecast for the quarter. To be clear, suppose a forecast for Q2 (April, May, June), made early March. I predict four months ahead (Mar, Apr, May and Jun) and subtract the stock of March (= forecast of the quarter). In this scenario, how to compare different forecasting models using a metric such as MAE and produce forecasting intervals?

  • $\begingroup$ I'm not following your question: What is preventing you from doing a quarterly forecast instead of a monthly one? $\endgroup$ – Skander H. Feb 4 at 23:34
  • $\begingroup$ @SkanderH. My dataset is small (65 months). So if I do such a prediction I will work with even a smaller amount of data. $\endgroup$ – rms Feb 7 at 12:30

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.