info on the MSVAR package and gretl 2025a
by Sven Schreiber
Dear (potential) users of the contributed MSVAR function package,
here's an update about the current situation. (BTW, MSVAR is about doing
frequentist Markov-Switching vector autoregressions, the package is
authored by myself, or to be more precise, it is a port of older Gauss
code by Anders Warne.)
Due to internal technical changes in the latest gretl 2025a release, the
current MSVAR version 0.22 will fail to run. This problem became
apparent in routine testing before the release, and a new MSVAR
candidate version 0.3 has already been submitted to the usual moderation
process; hopefully it will be public soon. If you're using current MSVAR
in real work (even though it is still somewhat work in progress), then
you should wait with the gretl update just a little longer.
cheers
sven
1 week
gretl 2025 and gretl4py released
by Allin Cottrell
See https://gretl.sourceforge.net/index.html
And for gretl4py (gretl for Python) see
https://gretl.sourceforge.net/gretl4py.html
Heads-up for Mac users: we've updated gretl from GTK2 to GTK3. So on
this occasion we recommend deleting your old Gretl.app before
installing 2025a.
Here are the gretl changes:
2025-03-20 version 2025a
- kalman filter: fixes for variance of smoothed disturbances
and some documentation updates
- "meantest" and "vartest" commands: support more options and
expand their documentation
- "system" command, iterated variants: allow more iterations
(now 250) before declaring non-convergence
- aggregate() function; speed improvement in the matrix case
- hyp2f1 function: enhance the documentation
- document the $pmanteau accessor for VARs
- GUI, preferences dialog: add a tab specifically for plot-
related settings
- Fix bug: potential crash in the GUI when trying to forecast
from an ARIMA model without a constant
- Fix bug: "dataset addobs n" with panel data (and without the
--panel-time option) was not respecting the documented
policy that 'n' refers to the number of cross-sectional
units to be added to the dataset
- Fix bug: "mpols" was not respecting the --simple-print
option, now it does
- Fix bug: the "join" command could overwrite some existing
data with missing values (contrary to the documentation) in
some cases
- Fix bug: the RESET test in the GUI was broken when the
"show all variants" option was selected (since 2024c)
- Function packaging: fix to allow a plot-checker function to
be private; documentation updates
- markdown parser: improve extraction of URLs
- minor code adjustments to comply with the C23 standard
- update sourceforge URLs to use https
- macOS builds: upgrade to GTK3
- macOS and 64-bit Windows: update gnuplot to version 6.0.2
- Windows 64-bit build: revert to use of MSVCRT rather then
UCRT, pending analysis of some problems with the latter
- Windows 64-bit build: update to OpenBLAS 0.3.29; fixes for
the Windows-10 theme
Allin Cottrell
1 week, 6 days
The function works fine, but the package doesn't!
by Paolo Chirico
Hi,
I have prepared a function to estimate stochastic volatility models with
three different QML procedures (which I attach with a datafile).
The first procedure is the classic QML, while the other two use
time-varying state space models, so they use update functions.
The script runs fine with all the procedures, Jack has already seen it.
However, after creating the package, only the first procedure works
well, while with the other two I get the message:
This function needs a dataset in place
The formula 'ERR = kfilter(&SVM)'
produced an error on execution
*** error in function svm_qm2, line 133
I think the problem concerns the update functions.
What is not working?
Thanks,
Paolo
--
Paolo Chirico
RU e Prof.Agg. di Statistica Economica
Università del Piemonte Orientale
Dip. di Giurisprudenza e Scienze Politiche,
Economiche e Sociali (DIGSPES)
Alessandria, Italia
--
Questa email è stata esaminata alla ricerca di virus dal software antivirus AVG.
www.avg.com
3 weeks, 2 days
Package updates (February 2025)
by Riccardo (Jack) Lucchetti
Dear all,
this message is to inform the community about the activity in our
function package repository: during the month of February 2025, 5
packages were updated to a new version:
"fdensity", by Artur Tarassow and me (factorized kernel density estimation)
"felogit", by me (Fixed-effects logit via Chamberlain's estimator)
"frontier", by Francesca Di Iorio and me (Stochastic Frontier Analysis)
"heatmap", by me (heatmap and contour plots)
"kmeans", by Artur Tarassow (K-means clustering)
"PairPlot", by Artur Tarassow and Allin Cottrell (Scatterplot matrix
with factor separation)
I you'd like to know more on these packages, or simply find out what's
new, the best way is to open the packages page
(https://gretl.sourceforge.net/cgi-bin/gretldata.cgi?opt=SHOW_FUNCS) and
click on the package name or on the [doc] tag if available.
Download and enjoy!
-------------------------------------------------------
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
-------------------------------------------------------
4 weeks, 1 day
gretl 2025a coming soon
by Cottrell, Allin
We expect to release version 2025a of gretl next week. If any
translators wish to submit updates this would be a good time to do so.
Allin Cottrell
4 weeks, 1 day
