[Gretl-users] Re: multi-step-ahead-forecasting
On Fri, 8 Jan 2021, Sven Schreiber wrote:
Am 08.01.2021 um 14:59 schrieb Burak Korkusuz:
> Is there a solution for that?
>
Well, see Allin's script a few postings back. Essentially using +1 as
the start and not +horizon.
In git, using +h (h > 1) for the start of a dynamic forecast will
now work. (We generate but do not report the required intervening
forecast values.) That's not yet in snapshots but will be soon.
Allin