Feature request for adding random variables
by andreas.karlsson＠ltv.se
Dear Professor Cottrell,
I have a small feature request for the Add > Random variables menu entry:
To make the GUI more intuitive and user friendly (so one does not always
have to look at the help files to understand the entries), please change
the Add > Random variables GUI to look like the one in Tools > P-value
finder, i.e. with tabs for the different distributions and text entry boxes
for mean, standard deviation, degrees of freedom, etc.
Also, please add the gamma distribution to this menu entry, so that one can
add random variables from the gamma distribution too.
Many thanks in advance.
Best regards and thanks for a great software,
Andreas Karlsson
18 years
Re: [Gretl-devel] observations limit
by andreas.karlsson＠ltv.se
cottrell(a)wfu.edu @ INTERNET skrev 2007-06-27 19:03:22 :
> I understand that current GTK contains a workaround that may fix
> this issue. The trouble is that GTK 2.8 and higher depend on cairo
> and don't run on win9X -- so gretl currently uses the GTK 2.6
> runtime. I think there are still people using gretl on win98, and
> I'm reluctant to try supporting two Windows packages for gretl.
Well, of course it is impossible to support two different Windows packages.
But on the other hand, improvements of gretl can not be halted just to
maintain support for obsolete operating systems.
I suggest that gretl 1.6.x continues to use the GTK 2.6 runtime, but that
gretl 1.7.x changes to a later version of GTK (I think that GTK 2.10 is the
latest version). And thus that gretl from version 1.7.x on no longer
supports win9x.
Regards,
Andreas
18 years
observations limit
by Sven Schreiber
Hi,
during a teaching session we got the error message "can't edit more than
1600 observations" (rough re-translation). This was unfortunate, can it
be generalized?
thanks,
sven
18 years
Reference Guide: missing commands.
by Hélio Guilherme
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Dear Allin and Jack,
I am translating the commands, and noticed that there is text but no
final output for these two "commands" (at least, I am only in the C letter):
compact
copy-formats
They are only GUI commands, and the help is relevant, but there are no
effective CLI commands, so the help is not produced. Should you create
dummy commands for this GUI exclusive type of commands.
Also the copy-formats text is very outdated and I think it should be
oriented to export-formats.
Thanks,
Hélio Guilherme
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFGfswdHvesJJ5UqvkRAn9zAKDalEminnwuK7TAPBvNaeIhnYxhwwCggS8D
kUx1WT1xzNUcaXmL9iMfDfY=
=iw9M
-----END PGP SIGNATURE-----
18 years
Ang. Re: [Gretl-devel] Feature request for the xtab cross tabulation function
by andreas.karlsson＠ltv.se
> On Mon, 11 Jun 2007, andreas.karlsson(a)ltv.se wrote:
>
> > Also, if it is not too hard, could you please add an option to save the
> > cross tabulation table that is produced by the xtab command as a
matrix?
>
> How much of this table do you think we should put into a matrix?
> Just the frequency counts? Allin.
I think that for a table with MxM cells of frequency counts it is most
useful to save it as an MxM matrix, and thus not include the row and column
marginal totals. These could easily be calculated anyway.
Best regards
Andreas
18 years
Feature request for the xtab cross tabulation function
by andreas.karlsson＠ltv.se
Dear Professor Cottrell,
Currently all options for the xtab command are not available from the GUI
at View > Cross tabulation.
Specifically, the options --row (display row percentages), --column
(display column percentages), and --zeros (display zero entries) are not
accessible from the GUI.
Could you please make these options available from the GUI too?
Also, if it is not too hard, could you please add an option to save the
cross tabulation table that is produced by the xtab command as a matrix?
Thirdly, if it is not too hard and does not take too much time to
implement, could you please add that Fisher's exact test is calculated and
displayed as default for 2x2 contingency tables, together with the usual
Pearson's chi-square test?
Many thanks in advance.
Best regards and thanks for a great software,
Andreas Karlsson
18 years
gretl crash
by Ignacio Diaz-Emparanza
1- Open data9-3.gdt
2- In the main window, If I select 8 series, right click and select "Time
series graph" all is ok and the graph is displayed. If I select 9 series,
gretl crashes. It doesn't matter which is the new series.
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU
Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.et.bs.ehu.es
18 years
function package GUI behavior
by Sven Schreiber
Hi,
I have noticed the following behavior for pre-selected values in the
function package GUI and have some questions or maybe suggestions:
* If a list argument doesn't have a defined default, gretl seems to
remember the last choice, which is very nice! Or is it merely the list
that was defined first? In any case I'm wondering whether it would be
possible to apply also this "memory" to list arguments _with_ defaults.
The default is IMHO intended to save typing in scripts, not to reset
your specification in the GUI every time. Something like the rule:
If there is a last saved spec. for the datafile, use it. If not, use the
default, if there is one. (But of course I understand if it's too
problematic to separate the GUI logic from the console use.)
* Same thing for other parameters; would it be possible to remember
them? I'm not sure what the logic is that gretl currently uses for the
preset values for example of matrix arguments that don't have a defined
default. The "oldest" in the session?
* What I find potentially dangerous is that gretl fills in some existing
object for example for a _return_ matrix. (Again, the oldest existing
return matrix in the session?) The default here should be empty IMHO.
* And something a little different: For matrix parameters, there is a
"new" button, while for list parameters, there is a "more" button. (I
know I'm partly responsible for that myself.) I would suggest some
unification: Either enable matrix editing from within the package GUI
like with lists, replacing "new" by "more" for matrices. Or _disable_
list editing from within the package GUI and implement that in gretl in
general. (Of course I would favor the last option, but since I am not
the one doing the work...)
Thanks for reading until the end ;-)
-sven
18 years
cointegration restrictions
by Sven Schreiber
Hi Allin and Jack,
since you mentioned that you're working on cointegration restrictions,
just a technical note on the switching algorithm for VECMs with alpha
and beta restrictions (Boswijk/Doornik section 4.4).
As could be expected, the starting values are very important. In
py4gretl_vecmrestrict <=0.9.4 (currently on the server) I simply used
the generically identified unrestricted estimates for alpha and beta
(and implicitly for the cov-matrix omega). However, that seemed to fail
quite often for the real-world cases I needed.
So today I tried out using generic starting values for alpha, beta, and
also omega, namely identity matrices (augmented by blocks of zeros to
get the right dimensions).
Although it's still very crude it seems to work much better! Of course,
one could add much more magic like comparing the results for different
starting values, scaling the data etc etc., but so far I don't need
that. The only difference with respect to results from PcGive is that my
standard errors are a bit higher, which I don't understand yet.
Apart from that I have also made a variant of the package which directly
uses the inputs G, h0, and H and is thus more flexible (but also less
user-friendly...). I'm not sure whether I should put that on the server,
too.
So much for that.
cheers,
sven
18 years
