Re: [Gretl-devel] [Gretl-users] multivariate lrvar?
On Mon, 16 Apr 2018, Sven Schreiber wrote:
Am 16.04.2018 um 13:12 schrieb Riccardo (Jack) Lucchetti:
> On Sun, 15 Apr 2018, Allin Cottrell wrote:
>>> Another difference is that lrcovar(X) corresponds to longrunvar(X, 1),
>>> assuming demeaned data. The question is whether to make the user
>>> responsible for demeaning or to offer it as an option.
>>
>> Easy enough to make it an option if that's desirable.
>
> Both things now in git. Sven could you check if the following is the
> behaviour you had in mind?
> A = lrcovar(X)
> B = lrcovar(X, 1)
You show that these two give the same result, thus 1 must be the default...
> C = lrcovar(X, 0)
> D = lrcovar(cdemean(X))
...and by that default, lrcovar(cdemean(X)) must be the same as
lrcovar(cdemean(X), 1). And since this gives the same result as lrcovar(X, 0)
I guess that "1" means: "don't touch it, it's already
demeaned", right?
So it seems you kept the semantics of the boolean switch, but you changed the
default from demeaned=0 to demeaned=1 I think. Is this deliberate?
Uh, sorry, I though I remembered that I had kept the semantics as the one
in the function you had posted originally, but maybe I got confused. I
don't know what the most intuitive syntax would be.
-------------------------------------------------------
Riccardo (Jack) Lucchetti
Dipartimento di Scienze Economiche e Sociali (DiSES)
Università Politecnica delle Marche
(formerly known as Università di Ancona)
r.lucchetti(a)univpm.it
http://www2.econ.univpm.it/servizi/hpp/lucchetti
-------------------------------------------------------