strptime() and Win vs Unix
by Riccardo (Jack) Lucchetti
Dear all,
a student of mine just let me know that he has problems with the yahoo_get
package on windows, when it comes to downloading data before 1970. This is
a consequence of the fact that our strptime() function behaves differently
on linux and windows. For example, the code
<hansl>
x = strptime("1964-05-11", "%Y-%m-%d")
</hansl>
returns generate a value for x that is -178074000 on linux but -1 on
windows. more generally, windows truncates any date prior to 1970 to -1.
Of course, writing a workaround for the affected package is relatively
simple, but shouldn't we address this inconsistency? Or this is not for
us, and we should just mention this fact in the docs? What do you think?
-------------------------------------------------------
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
-------------------------------------------------------
8 months
How could ESS data be weighted in Gretl software package?
by Μπεκιάρης Ευάγγελος
Please post to the "Gretl-users" mailing list the following question:
I have easily weighted data in SPSS in the past, and now I would like to weight European Social Survey (ESS) data in the GRETL software package, if possible, but I don't know how. I would be most grateful if you advised me.
The syntax I use in SPSS for this purpose is as follows: WEIGHT BY analysisweight .
Best Regards,Evangelos Bekiaris PhD candidate University of Peloponnese
8 months
kmeans v0.2
by Artur T.
Hi all,
since yesterday version 0.2 of the "kmeans" package for doing
unsupervised learning is available:
http://gretl.sourceforge.net/current_fnfiles/kmeans.gfn
This version fixes a bug that arises if the sample range is restricted
and you're trying to coerce a column vector that's not the full length
of the dataset into a series on adding it to a bundle.
Enjoy the package and please report any eventual bugs.
Best,
Artur
8 months, 1 week
Wald test save results
by Filipe R. Costa
Hi all,
Is it possible to capture the results from Wald tests linear restrictions similarly to what we can do with $model into a bundle?
Let me use an example file from Gretl:
<hansl>
open AWM18.gdt
# model 1
ols YER const PCR GCR
bundle bundle1 = $model
# Apply the linear restriction
restrict
b[2]-b[3]=0
end restrict
</hansl>
I'm interested in the test statistic, p-value and eventually degrees of freedom. Any very simple way to retrieve them, without computing manually from bundle1 results?
Thanks in advance,
Filipe Costa
8 months, 1 week
GARCH News Impact Curve
by Theodoros Panagiotidis
Dear all,
has anybody tried to produce news impact curve in from the GARCH option
(GUI) option in gretl and from the GARCH variants (gig 2.24)?
Is the option residuals from the GARCH option in GUI the same as the
standardised residuals in the GARCH variants gig?
regards
Theo
8 months, 2 weeks
import problem with quotes in text data files
by Sven Schreiber
Hi,
attached is a minimal example file. There are two data columns (no
observation column), and because the blank (space) is the column
separator, the column headings are enclosed in double quotes at the source.
(I did not make up this format.)
Gretl imports the data alright, but kind of ignores the double quotes,
treating the first two space-separated words as the names, which gives
totally wrong headings after the first one. The only workaround that I
can see is to manually replace the interior blanks with underscores or
something like that?
Note that during the import gretl tells me that it "treats double quotes
as significant", whatever that means.
Of course gretl cannot get every weird text file format right, but at
the same time AFAIK it is impossible to provide some helpful hints how
to interpret the contents.
thanks
sven
8 months, 2 weeks
Forecasts (from a single equation) with (G)ARCH effects
by Sven Schreiber
Hi again,
I'm wondering what's the status quo about forecast error/confidence
bands in equations with any kind of ARCH effects, say a simple ARCH(1)
or GARCH(1,1). In particular:
- When I estimate the model natively in gretl (Model - Univariate Time
Series - GARCH), I can call up the standard forecast dialog from the
model window (Analysis - Forecast menu), and I get the usual plot with
confidence bands. But I'm not sure whether the confidence bands actually
incorporate the conditional variance from the GARCH part of the model
(from the variance equation). I cannot easily check against a plain OLS
model without GARCH because I also get slightly different coefficients
in the mean equation there. I looked at the built-in help and at the
manual, but it seemed silent on this particular case.
- If it so happens that in the native forecasts the GARCH effects are
not actually taken into account, then there still is the "gig" package.
Its help has a section about forecasting, but that only seems to deal
with forecasting the variance itself. Now, I believe I could use this
variance forecast from gig to combine that with the point forecasts from
the mean equation, in order to construct forecast error bands around the
mean equation. Two questions here: Do you agree that that's possible?
And secondly, would there be an easier way than combining it manually?
Hope I made myself clear. My target is always the forecast of the first
moment (the mean) and its uncertainty, the variance is only of indirect
interest.
thanks
sven
8 months, 2 weeks
No date information in b-g.gdt example dataset
by Sven Schreiber
Hi all,
with a fairly recent snapshot I've loaded the shipped example dataset
b-g.gdt, which gretl at the bottom announces as 5-day daily and
incomplete. OK - does "incomplete" mean gappy here? And there is no
information about the starting or ending dates, and possibly not on
where the gaps are, either. (Not sure about this one because it's a lot
of observations.)
Does anybody know more about this data?
thanks
sven
8 months, 2 weeks
Data with gaps
by Alecos Papadopoulos
I fished the following from the cross-validated Q&A forum
https://stats.stackexchange.com/q/581717/28746
<<
This is my first research using Gretl. I have a panel of 197 people for
66 months stacked as time series. One person has values for the months
of January, February, Mars, April, November, December. For the summer
months of May, June, July, August, October and September there is no
data available. so my data goes from 2005-11 to 2016-04 with 6 months
missing every year. It is not surprising, that Gretl does not recognize,
that the summer months are missing. It strings the data together and
recognizes my data from 2005-11 to 2011-04. Does any of you have an idea
how to solve that problem and make Gretl realize i only have 6 months
per year?
My first idea was to include missing values for the summer months. Is
this possible in Gretl or do I need to do it in Excel?
I already checked in the manual and various internet plattforms, but
could not find a solution. So I am greatfull for any suggestion.
>>
Can we help this person?
--
Alecos Papadopoulos PhD
Affiliate Researcher
Dpt of Economics, Athens University of Economics and Business
Foundation for Economic and Industrial Research (IOBE)
web: alecospapadopoulos.wordpress.com/
scholar:https://g.co/kgs/BqH2YU
8 months, 2 weeks
cubic splines, fractional polynomials, and more
by Riccardo (Jack) Lucchetti
Hi all,
I'd just like to tell people that we have a new function package (by yours
truly) called "basis_functions", that can be used for easily incorporating
nonlinearity in your models; it also includes a handy function for
generating Fourier harmonics.
You can find it in the usual place:
http://gretl.sourceforge.net/cgi-bin/gretldata.cgi?opt=SHOW_FUNCS
Of course, comments are welcome.
-------------------------------------------------------
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
-------------------------------------------------------
8 months, 2 weeks