9:43 p.m.
Hi, Allin;
Yes, with the spectrum there is no problem, the problem appears when you
select the periodogram (/Variable/Spectrum/ sample periodogram)!
Cheers
Javi
10:54 a.m.
Am 01.07.2010 02:16, schrieb Allin Cottrell:
On Wed, 30 Jun 2010, [iso-8859-1] Javier Garc�a wrote:
> Yes, with the spectrum there is no problem, the problem appears when you
> select the periodogram (/Variable/Spectrum/ sample periodogram)!
Ah, OK, I see what you mean. This is noew fixed in CVS.
Hold on a sec -- that's what the other submenu entry in
/Variable/Spectrum below "periodogram" is and has always been for, no?
The periodogram itself of course doesn't have any smoothing, so no
bandwidth. Though I guess one could argue to substitute the two-item
submenu under /Variable/Spectrum with a single dialog offering the
choice instead.
-sven
7:36 p.m.
On Thu, 1 Jul 2010, Sven Schreiber wrote:
Am 01.07.2010 02:16, schrieb Allin Cottrell:
>
> On Wed, 30 Jun 2010, [iso-8859-1] Javier Garc�a wrote:
>
>> Yes, with the spectrum there is no problem, the problem appears when you
>> select the periodogram (/Variable/Spectrum/ sample periodogram)!
>
> Ah, OK, I see what you mean. This is noew fixed in CVS.
Hold on a sec -- that's what the other submenu entry in
/Variable/Spectrum below "periodogram" is and has always been for, no?
The periodogram itself of course doesn't have any smoothing, so no
bandwidth. Though I guess one could argue to substitute the two-item
submenu under /Variable/Spectrum with a single dialog offering the
choice instead.
The periodogram doesn't have any smoothing, no -- and that fooled
me into thinking that the periodogram dialog should not have a
bandwidth spinner. So I had removed it, and that's what Javier was
complaining about.
The reason why it _should_ have an adjuster is to govern the order
for the fractional integration tests, something I had forgotten
about (though it's in the manual). But perhaps the label for the
spinner should be different in the two cases (sample periodgram /
Bartlett thingy)?
Allin
7:41 p.m.
Am 01.07.2010 21:36, schrieb Allin Cottrell:
The reason why it _should_ have an adjuster is to govern the order
for the fractional integration tests, something I had forgotten
about (though it's in the manual). But perhaps the label for the
spinner should be different in the two cases (sample periodgram /
Bartlett thingy)?
I think the proper solution would be to separate the fractional
integration tests from the estimate of the periodogram. Sure they're
related, but so is the periodogram and the spectrum, and they're
separate. Right now the fractional integration tests are actually quite
hidden, for example I found out about them basically by chance.
cheers,
sven
9:04 p.m.
On Thu, 1 Jul 2010, Sven Schreiber wrote:
Am 01.07.2010 21:36, schrieb Allin Cottrell:
> The reason why it _should_ have an adjuster is to govern the order
> for the fractional integration tests, something I had forgotten
> about (though it's in the manual). But perhaps the label for the
> spinner should be different in the two cases (sample periodgram /
> Bartlett thingy)?
I think the proper solution would be to separate the fractional
integration tests from the estimate of the periodogram. Sure they're
related, but so is the periodogram and the spectrum, and they're
separate. Right now the fractional integration tests are actually quite
hidden, for example I found out about them basically by chance.
I think you're right. I'm trying a different layout in CVS and
snapshots. What do you think of it?
Allin
5042
days inactive
5043
days old
5 comments
3 participants
participants (3)
-
Allin Cottrell -
Javier García -
Sven Schreiber