ACF and documentation of Bartlett standard errors
by Sven Schreiber
Hi,
I think the choice of Bartlett standard errors for the correlogram in
the GUI is relatively recent (not sure), while the corresponding
--bartlett option for the corrgm command has been there for a long time.
In any case, I was made aware of the special null hypothesis underlying
the Bartlett variant: namely that for the interval at lag k the null
hypothesis is an MA(k-1). This is not a problem (of course), but the
standard confidence intervals come from white noise as the null
hypothesis. So the choice of Bartlett standard errors in this case isn't
just some robustification or small-sample thing (as opposed to a
Bartlett correction in other contexts, for example), but involves a
different view.
So far so good, but I haven't found any mention of what the Bartlett
errors actually mean. I have checked the GUI help text (for
"Correlogram"), the documentation of the corrgm command, and the user
guide. Actually, I was a bit surprised that the user guide and
specifically the univariate time series chapter does not mention the
ACF/PACF/correlogram at all. (Searching for "corrgm" or "ACF" doesn't
show anything, and "correlogram" only shows up in a different context as
cross-correlogram.) This is not a criticism, but just an observation --
maybe there are places I have overlooked.
I think wherever these Bartlett standard errors are mentioned as an
option, the different null hypothesis should be made explicit. I also
think it would be good to put the word "Bartlett" into the ACF plot if
it's chosen.
I'm posting this here for discussion, maybe I'm misunderstanding the
intention or assumption for those Bartlett standard errors, since I
haven't checked the source code (yet).
thanks
sven
2 months, 1 week
ui-maker and no_const
by Sven Schreiber
Hi,
is the "no_const" specification for the ui-maker apparatus for a
function package GUI really working as intended? I seem to have a case
where I have a ui-maker function that requests no_const for a list
argument X, but I can execute the function with X = const alright. BTW,
X is set to have a "null" default, if that matters. Or does no_const
really mean that any constant term is removed quietly?
I don't have time to give a minimal example right now, but I could do so
on Friday or so, if necessary.
thanks
sven
4 months
intended effect of 'set force_decpoint off' ?
by Sven Schreiber
Hi,
I'm observing the following behavior on a German system, but running
gretl without the local dec sep setting, i.e. not using the comma.
<gretl-console>
? = 1.6
1.6
? set force_decpoint off
? = 1.6
1,6
</gretl-console>
The first thing is OK, gretl showing in response that it's using the
point (period) for display. After setting force_decpoint to off, gretl
is then using the comma. The background is that this is screwing up some
code.
I'm having a deja-vu about this: But I believe that our conclusion in a
previous discussion was that if I don't want to _force_ the decimal
point any longer, then I want to revert to the current setting as
governed by the user preferences, _not_ to use the default language
setting. In my case, continue to use the point.
So, has this behavior crept back in? This is with the snapshot from Aug
5th (standard, GTK2) on Windows.
Oh, and here's another interesting printout, coming from gretlcli.exe:
<gretlcli>
? =1.6
1.6
? set force_decpoint off
? = 1.6
1
</gretlcli>
thanks
sven
4 months
Re: [Gretl-users] Re: new gretl snapshots available
by Sven Schreiber
Am 23.07.2024 um 17:11 schrieb Cottrell, Allin:
> On Tue, Jul 23, 2024 at 8:49 AM Sven Schreiber
> <sven.schreiber(a)fu-berlin.de> wrote:
>> Am 22.07.2024 um 20:55 schrieb Cottrell, Allin:
>>
>>
>> Thanks, Allin. A first bit of feedback: When I have language preferences set to "Automatic" (on a Win11 system set to German) and change the _editor_ theme to something else, on restart then gretl seems to come up in a language mish-mash: The menus are in English, but other labels are in German. But another restart apparently solves that; not sure whether another change of settings was required for that.
> I'll try some testing on that.
(moving this to the devel-list; all this is with the experimental
GTK3-based Windows snapshot.)
I have just installed the latest incarnation from Aug 3rd. This is here
on Win10 now. On the attached screenshot you can see that on the left of
the main window some text and other stuff is not visible, it's beyond
the window virtual border. This seems to be due to the workdir path; if
that is shorter, the stuff becomes visible again. I believe that under
GTK2 the path display is automatically shortened. This problem
disappears when a dataset (with a shorter name than the German message
"Keine Datendatei geladen") is loaded; or when the console on the right
is narrower, such that the left-hand side has more space.
Fonts are very sharp on this hi-res laptop screen, very nice!
Honoring the theme setting and/or the language setting still sometimes
seems to require one extra restart. For example, the screenshot shows
Adawaita (I believe), while the theme setting says Windows-10.
BTW, on Win11 the window edges are more rounded again, so there I
actually even prefer the old Win-7 theme or even Adawaita over the
Windows-10 theme. I guess a usable Windows-11 theme doesn't exist yet
for GTK3?
thanks
sven
4 months, 1 week