On Fri, 24 Sep 2021, José Henriques wrote:
I think if I open the file directly on Gretl it may be the best
option.
My doubt is that Gretl doesn't read some part of the data because
it's already in the file and when preparing the forecasting model,
I must enter new information in empty cells in order that the
programme processes the new information and not the one that was
already in the file (that was opened directly).
I will try it both ways (open it directly and the CSV file) to see
if it works.
If I understand the context correctly, here's what I recommend.
* Construct your complete dataset, including observations intended
for forecasting, in whatever way you're most comfortable with.
* Load the dataset into gretl (from CSV, Excel or whatever).
* Set the sample range to include just the observations you wish to
use for estimation, and estimate your model.
* Now (if there are extra data available) gretl will give you the
option of making an out-of-sample forecast.
Allin Cottrell