1
$\begingroup$

I'm looking at 30 years of rainfall data, working with the daily average rainfall per month (That sounds weird... it's the total rainfall in the month, divided by number of days in the month).

Anyway, after seasonal differencing it looks to fit to a SMA(1) process relatively well, but the forecast Minitab gives includes negative rainfalls. Is there a way to define that only values zero and up are appropriate for this forecast? Or perhaps something else I'm meant to have done with my data first?

$\endgroup$
2
  • $\begingroup$ Side question @gung, for future reference, is it against some rules here to add niceties to a question? $\endgroup$
    – MissGeek
    Jun 6 '15 at 11:35
  • 1
    $\begingroup$ Eg, "thanks"? Yes, we ask that you do not do that. The point here isn't so much to help you per se, as it is to build up a permanent repository of high quality statistical information in the form of questions & answers. "Thanks" etc, distracts from that ever so slightly. When someone answers, you can thank them by accepting their answer. $\endgroup$ Jun 6 '15 at 13:07
1
$\begingroup$

Using SMA(1) there is no option to restrict like the way you want. You may want to take the log of the variable to avoid negative values and do the forecasts, and back-transform

Please see here for further reference.

$\endgroup$

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.