Change graph resolution
by Artur Tarassow
Dear all,
I would like to change the resolution of a graph. See the following example
<hansl>
open denmark.gdt -q
list lplot = LRM LRY
P1 <- plot lplot
options with-lines time-series
literal set terminal pngcairo size 720,320 enhanced font 'Verdana,12'
end plot --output="display"
</hansl>
which doesn't adapt the resolution accordingly. However, doing the
following steps creates the correctly sized figure:
1) right click on object P1 in the icon view --> Edit plot commands
2) Add the following to line 14: set terminal pngcairo size 720,320
enhanced font 'Verdana,12'
3) Save (Cmd+s) and run (Cmd+r) within the window, and close the window.
4) re-open P1
Is there a reason why setting the resolution is not allowed? I am
surprised to see that the "set terminal ..." line does not appear in the
written gnuplot-script even though it's within the plot-block.
Best,
Artur
7 years, 3 months
Unable to save save a record of commands issued.
by John C Frain
I recently completed a clean reinstall of 64-bit Windows 10 1709 on a
laptop. As part of the process I updated gretl to the current snapshot.
At the end of a session I am prompted to save a record of commands issued
which I wish to do. When I try to do so I get boxes with the following
messages
---------------------------------------------
gretl error
Failed to open file 'C:\Users\frain\AppData\Roaming\gretl\prntmp.NvmQJo':
Permission denied
Close
---------------------------------------------------
and when I close this box
---------------------------------------------------
gretl error
Error attempting to open file
Close
-----------------------------------------------------
I also used windows update to update a tablet computer to 32-bit Windows 10
1709. After the update 32-bit gretl2017d appeared to run without the
problem. I downloaded and installed the snapshot version yesterday. The
session record problem now appeared on the 32-bit tablet. I uninstalled
the snapshot version and reinstalled the 2107d version but now the problem
is in the 2107d version.
I have also uninstalled the 64-bit version (using the supplied uninstall
program) and reinstalled the 64-bit 2107d version and still have the
problem.
I should say that there is no obvious problem with the
'C:\Users\frain\AppData\Roaming\gretl\' directory. The file
'prntmp.NvmQJo' has only the archive attribute set. Also I was able to
copy the messages to a text file in the directory. I have been unable to
copy or open the 'prntmp.NvmQJo' and when I close gretl the file
vanishes. I presume that the file is a temporary file containing the
instructions issued during the session.
John C Frain
3 Aranleigh Park
Rathfarnham
Dublin 14
Ireland
www.tcd.ie/Economics/staff/frainj/home.html
mailto:frainj@tcd.ie
mailto:frainj@gmail.com
7 years, 4 months
Boxplot graphic
by juergen.malitte＠t-online.de
Hi all,
I think, there’s a problem concerning the processing of a boxplot
graphic. Choose any dataset-variable and generate a boxplot grafic (click
right mouse button and choose „boxplot“). In the boxplot window click the
right mouse button and choose „bearbeiten“. Now one have the possibility
to add some text („Bezeichnungen“) or to add arrows/lines („pfeile“) for
example. If you do so, the plot of the box will be compressed to a small
line.
With greetings
Jürgen Malitte
7 years, 4 months
Re: [Gretl-users] Message "data error" in MLE
by Ioannis A. Venetis
One guess is trouble will the log expression and/or exponentials etc.
Try to simplify the likelihood function.
For example -2*log(th)+log( exp(.+.)*B1 + (resM+su)*B2 + sr*B3 )
Yiannis
7 years, 4 months
Message "data error" in MLE
by Alecos Papadopoulos
Good afternoon.
I keep getting a "Data error" message (and nothing else) in an ML
estimation with a non-standard density (which has been checked of being
a density , i.e. non-negative and integrating to unity).
I have used the same data set with other densities without problems.
There are two regressors and a constant. Sample size is 570.
What such an error message may signal?
The script is
mle logl = check ? log((2/th^2)*exp(sr^2/(2*su^2))*exp(resM/su)*B1 +
(2/th^2)*(resM+su)*B2 + (2/th^2)*sr*B3) :NA
series resM = Depv - Regrs*kVec
series B1 = su*cnorm(-resM/sr - sr/su) - (th+su)*cnorm(-resM/sr -
sr/su + th/sr)
series B2 = cnorm(resM/sr) - cnorm((resM-th)/sr)
series B3 = dnorm(resM/sr) - dnorm((resM-th)/sr)
scalar check = (sr>0) && (su>0) && (th>0)
params kVec sr su th
end mle --verbose
--
Alecos Papadopoulos
PhD Candidate
Athens University of Economics and Business, Greece
School of Economic Sciences
Department of Economics
https://alecospapadopoulos.wordpress.com/
7 years, 4 months
