Re: [Gretl-users] Re: Gretl-users Digest, Vol 24, Issue 2
On Tue, April 4, 2006 09:47, Ignacio Díaz-Emparanza wrote:
El Martes, 4 de Abril de 2006 09:21, John Paravantis PhD escribió:
> Some points on the feedback appearing below:
>
> - SPSS, Minitab etc all successfully estimate these ARIMA models and it
> is left up to the user to decide how much trust should be placed with
> the estimates. Gretl fails to provide an estimate and I find this annoying.
I agree with you. It is better to have a bad estimation than have no
estimation. I suppose we need a better optimization routine.
I find it annoying when you get an "estimate" which has no statistical
properties whatsoever and just happens to be the point in the parameter spece
you happened to be at when you reached the maximum number of iterations. I
understand that if you want to sell your package, you may want to avoid giving
the user the impression that your package doesn't work, but we don't need to
do that. In case you want to know what's going on, that's what the --verbose
switch is for.
>
> How about making the number of iterations and tolerance available in the
> dialog box version of ARIMA?
I also agree that this should be available in the GUI, but I do not know
whether is better here or in the "preferences" menu.
IMO, we should have some graphical equivalent of the "set" command.
Unfortunately, it's been difficult to work on gretl for a few days because of
a hardware failure at sourceforge. When we get back in action, that's
something we can think about.
Riccardo "Jack" Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona