Bug for XY scatterplot
by andreas.karlsson＠ltv.se
A bug for XY scatterplots in gretl:
1. Plot a XY scatterplot from View > Graph specified vars > X-Y scatter...
2. Click on the graph to obtain the pop-up menu and choose Edit > Main >
Fitted line > quadratic > OK
3. Click on the graph to obtain the pop-up menu: The "OLS estimates" menu
entry has disappeared.
Best regards
Andreas Karlsson
18 years
py4gretl_vecm-0.9.1
by Sven Schreiber
Hi,
I have just uploaded version 0.9.1 of py4gretl_vecm to the gretl
function package server. It includes a little bugfix so that the temp
files should really be deleted on Windows (at least it now does so for me).
How can I remove the obsolete version 0.9.0 from the server?
Since the py4gretl packages now seem to work also on Windows at least
for me, I have also uploaded py4gretl_vecdecomp and py4gretl_vecrestrict
to the package server (both also at version 0.9.1). See the readme on
http://www.wiwi.uni-frankfurt.de/profs/nautz/schreibersoftware.html for
mor information. (Actually I just noticed that there is probably a typo
in the URL given in the help texts of the functions, sorry. The one
above should work.)
All packages still need testing, that's why the version numbers are
still less than 1.0.
I hope this is useful at least for some people,
Sven
(Allin: I still need to restart gretl --current win snapshot-- to make a
new package, after I have already packaged other stuff, because
otherwise the function list is empty or incomplete. Just fyi.)
18 years
command gnuplot
by Marco Marini
Hi everybody! First time for me in this mailing list, I'm
a new user of Gretl and I am more and more glad to work with it!
When I make a graph of a time series from the GUI, I obtain
correctly a line between observations. Looking at the commands log, I see
the command gnuplot <name> --time-series.
If you type the same command on the terminal, the graph does not show
lines but only the + marker. Is that a mistake?
Thanks to all and regards
Marco
*Marco Marini*
ISTAT - National Statistical Institute
Quarterly National Accounts
Via Agostino Depretis 74/b
00184 Rome
+39.06.4673.3209
marco.marini(a)istat.it
www.istat.it
18 years
Ang. Re: [Gretl-users] Check for idempotent matrix in gretl
by andreas.karlsson＠ltv.se
> On Fri, April 13, 2007 23:17, Riccardo (Jack) Lucchetti wrote:
[...]
> Thinking a bit more about it, I thought it would be way more economical,
> from a computational viewpoint, to decide whether a matrix is idempotent
> or not simply by a multiplication check, because matrix multiplication is
> much cheaper than the eigenproblem. But, may I ask what this check is
> for?
Idempotent matrices play an essential role for the conditions under which a
quadratic form in normal variates has a chisquared distribution. See
Chapter 10.4 in James R. Schott (2005), Matrix Analysis for Statistics,
2nd. ed. Hoboken, NJ: Wiley.
On checking the idempoteny, I just thought of using a multiplication check.
Best regards,
Andreas
18 years
Check for idempotent matrix in gretl
by Andreas Karlsson
Hi,
Could you please add a check for idempotency in matrix? So that for a matrix,
when one chooses Prpoerties > View the output tells if the matrix is
idempotent, just as it now tells if it is e.g. square, symmetric or positive
definite.
Thanks in advance.
Best regards,
Andreas K.
18 years
gretl for Google's summer of code next year
by Talha Yalta
Google has just announced accepted projects for their yearly "summer
of code" event. This year they accepted 900 students (from 6200
applicants), who will be working over the summer to improve 140 FLOSS
projects. Students receive $4,500 for completed projects, and the
mentoring organizations receive $500 for each project. Google stands
to pay out up to $4.5 million if all of the projects reach completion.
The list of this year's accepted projects can be found here:
http://code.google.com/soc/
Among various projects, I notice one regarding integration of R into
openoffice Calc for example. I think it might be a good idea for gretl
to apply and participate in this event next year with several projects
that will improve various aspects of the program. For example, I
remember Allin looking for volunteers some time ago for improving the
preferences windows so that more options can be set using the GUI
instead of the console commands etc. Something to think about.
Cheers
Talha
--
"Money won't buy happiness, but it will pay the salaries of a large
research staff to study the problem." - Bill Vaughan
--
18 years
model table news
by Allin Cottrell
In response to Talha's suggestions, gretl CVS and the current
Windows snapshot incorporate these changes in relation to the
"model table" feature:
* You can add up to 12 models to the table.
* In the GUI, the text in the model table window does not wrap,
instead you get a horizontal scrollbar if the table is too wide.
* In TeX mode, if there are more than 6 models the table is put
into landscape.
* The default formatting spec for TeX material in gretl uses the
"11pt" option for the article document class, which IMO makes for
easier reading on-screen. But for the model table we now leave
the font size at the TeX default of 10-point, to increase the
chances of big tables fitting on the page.
* The TeX model table now uses the "longtable" package, so if you
have a table with lots of regressors, the table should break
gracefully over the page.
TODO: Not all of these changes (in regard to TeX) are handled
properly (yet) if you're using a customized gretl TeX preamble.
But that shouldn't be too hard to fix.
Allin.
18 years
new Windows snapshot
by Allin Cottrell
Perhaps of interest to anyone experiencing the problem on exit
that Talha mentioned. There's a new Windows snapshot at the usual
place:
http://ricardo.ecn.wfu.edu/pub/gretl/gretl_install.exe
This has some modifications to the "exit" code, although since I
haven't identified the problem with any confidence I don't know if
it will help.
This build also incorporates Andreas's suggestion of showing the
median of the dependent variable for LAD models.
Allin.
18 years
leverage graph
by ab.news
Hi everybody!
While executing the leverage command from the GUI (Model
->Tests -> Influential observations)
the twin graphs produced don't really seem like gretl's
habitual graphs.
I used the following script for quarterly data from 1950:1 to
2000:4
open "C:\Program Files\gretl\data\greene\greene5_1.gdt"
ols realgdp const pop
/* commands pertaining to model 1 */
leverage
and both graphs:
- don't show the X-axis left-margin;
- the X-axis is expanded far behind the last observation
(until 2010 ?)
Is there something wrong or I'm missing something ?
Thanks from Artur
GRETL'S GREAT
Créez votre adresse électronique prenom.nom(a)laposte.net
1 Go d'espace de stockage, anti-spam et anti-virus intégrés.
18 years
TeX export issues
by Talha Yalta
I have been using the TeX export functionality of gretl, which is very
useful. I also noticed some problems. I don't know if these are bugs
but I wanted to report nonetheless:
1- In the TeX file of the model tables I see the "\footnotesize{}"
command before each item in each cell, which makes the table difficult
to read and touch up in LaTeX. Wouldn't it be better if there was
\begin{footnotesize} and \end{footnotesize} at the beginning and the
end of the table?
2- In some cells I have "$\,\,$" or $\,$. I am not sure what these are
for (I just erease these).
3- Instead of the "$-$0.7971$^{**}$" format, wouldn't it be better for
all the cells containing numbers to just be in the formula mode such
as $-0.7971^{**}$ ?
4- After experimenting a little bit, I now see that 6 models is what
can be fit in a portrait page. Still, could you please make it 8 or
even 12? I know 6 will be enough for most purposes however:
a)- there might be need for more than 6 in some cases (right now I need 8)
b)- Even though more than 6 is difficult to fit in a portrait page, it
is possible to choose a landscpe page later (would be even better if
there was an option for this in TeX export options).
c)- Again the TeX code can be tweaked later so that the table is in
footnotesize or scriptsize and it fits in a portrait page even.
5- Request: I don't know how difficult this is to implement but
exporting a table in a way where all &'s are vertically alligned would
make it easier to read and edit in TeX later. Example:
Debt / wealth ratio & 0&015 & 0&023 & -0&000 \\
& (2&490)** & (5&273)**&
(-0&028) \\
Dummy if head is white & 0&461 & -0&450 & 0&057 \\
& (-2&424)**& (-3&313)**&
(0&208) \\
Cheers
Talha
--
"Money won't buy happiness, but it will pay the salaries of a large
research staff to study the problem." - Bill Vaughan
--
18 years
