Package updates (December 2024)
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 December 2024, 3
packages were updated to a new version:
"graphlasso", by Sven Schreiber (graphical Lasso --- l1 shrinkage for
covariance matrices)
"lomackinlay", by Allin Cottrell and Sven Schreiber (Lo-MacKinlay
variance ratio test for determining if a time series is a random walk
process)
"BACE", by Marcin Błażejowski and Jacek Kwiatkowski (Bayesian Averaging
of Classical Estimates)
I'll reserve a special mention for the "graphlasso", which now uses a
new internal function to speed up computation.
Sorry if this message comes relatively late in the month, but I was away ;)
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
-------------------------------------------------------
1 week, 6 days
ARIMA on log transformed data
by dbrilakis＠yahoo.gr
Hi, I found that my data become stationary (after differentiate) the log data with best ARIMA(p,d,q) How do I rebuilt the ARIMA forecast to the origina scale before log?
Tanks in advance
2 weeks, 3 days
Assistance Needed on Scaling Impulse Response in SVEC Model Using "gpbuild"
by Olasehinde Timmy
Dear All,
I hope you are doing well.
I am working with the *SVEC model* and using the *"gpbuild"* syntax to
generate impulse response functions. However, I am having difficulty
adjusting the scale to make the impulse response plots more visible and
readable. I would appreciate any guidance on how to modify the settings or
apply any specific syntax that can enhance the visualization.
If anyone has experience with this or knows of a useful reference, I would
be grateful for your assistance.
Looking forward to your suggestions.
Best regards,
Timmy
3 months, 2 weeks
Set number of max. iterations for 3SLS
by Artur T.
Hi all,
I am estimating a system by means of 3SLS. The default max. number of
iterations is set to 100. The estimator does not converge for my
specification, and I would like to increase the max. number.
However, I can't find the parameter option for the set command to
control the max. number of iterations.
For illustration: Here as an example for which the estimator converges
after 39 iterations. I am trying to set the max. number to 10 using
three different "maxiter" options but none of them is respcted.
<hansl>
set verbose off
open klein.gdt
series W = Wp + Wg
series A = t + (1918 - 1931)
series K1 = K(-1)
# set the model up as a system
"Klein Model 1" <- system
equation C 0 P P(-1) W
equation I 0 P P(-1) K1
equation Wp 0 X X(-1) A
end system
# 39 iterations until convergence
estimate "Klein Model 1" method=3sls --iterate
# Set max. 10 iterations --> Should not converge
set bfgs_maxiter 10
set bhhh_maxiter 10
set gmm_maxiter 10
estimate "Klein Model 1" method=3sls --iterate
</hansl>
Thanks,
Artur
3 months, 3 weeks
Add DAta -Data entry Gretl2024d
by Brian Revell
There is still an issue with the add data entry function If a double return
does not register (as is possible when rapidly inputting lots of data
quickly) any simple and move the cursor to the next cell ( ie 2 entries are
typed inadvertently into the same cell, Gretl crashes. So, if the user
input error occurs during a long series of data all previous inputs are
lost. In addition, if the user has not saved any other operation outputs or
analysis before the data entry error occurs, all that is lost too. I
would humbly suggest that any data error entry in the Add Data function
calls up a data entry error warning rather than crashing Gretl.
Brian
3 months, 3 weeks
K means clustering algorithm
by Brian Revell
HI
the package gives an error message when attempting to install Arthur's
package
viz. logging.zip Not found on server.
Brian J Revell
PS Happy New Year
3 months, 4 weeks
Inputting new data
by Brian Revell
I seem to be experiencing crashouts of gGretl when adding a new variable or
revising existing data.
After entering an observation and pressing return, the cursor moves to the
next observation cell. However on typing in the next observation, nothing
appears in the cell and Gretl crashes out. This should not happen and is
extremely frustrating. I have never experienced this on earlier versions
(current version 2024d).
The only work around solution is to physically the select the next cell by
clicking on it, and entering the observation data. Harly a streamlined
input process.
Please check.
Given the problem I have had loading k-means clustering package (still not
solved) , I am wondering if I shoudld simply redownload an earlier version
that 2024d?
Advice welcomed.
Brian
Brian J Revell
Professor Emeritus (Agricultural Economics)
Tel: home 01952 728153 Mobile 07976 538712
Address: Orchard Croft, Vineyard Rd, Homer, Much Wenlock TF!3 6NG
Alt. Email bjrevell(a)harper-adams.ac.uk
3 months, 4 weeks
New Version 0.4 of the kmeans package
by Artur T.
Dear Gretl users,
I'm pleased to announce version 0.4 of the kmeans package. This update
includes several important improvements:
1. A new function `kmeans_screeplot()` has been added, which helps in
determining the optimal number of clusters through visual analysis.
2. The `kmeans_plot()` function has been updated to use Gretl's default
values for pointsize and fontsize parameters, providing better
consistency with the overall Gretl interface.
3. We've fixed a bug related to distance measures - the function now
correctly uses the specified distance measure when called through the
GUI, rather than defaulting to the standard measure.
You can update the package through Gretl's package manager. More details
here:
https://gretl.sourceforge.net/current_fnfiles/kmeans.gfn
Best regards,
Artur
4 months, 1 week
Contributed package buys_ballot will be retired
by Sven Schreiber
Hi everybody,
this is to pre-announce that the veteran contributed function package
"buys_ballot" by Ignacio Díaz-Emparanza (and Jack Lucchetti, last
updated in 2018) for seasonal plotting will be retired soon from the
public package server. The reason is that its functionality and a little
more is offered by the newer package "season_plot" by Artur Tarassow.
This course of action has been discussed and coordinated with the
package authors.
One restriction to note is that "season_plot" for technical reasons
requires at least gretl 2024b, which is relatively recent. So we are
going to wait for the next gretl release 2024d (which is expected soon)
before retiring the other package. As always, you can continue to use
the package and also store a local copy of "buys_ballot" for further
use. It's probably also going to be archived (@Artur?).
If you notice any problems in your usage with the newer package
"season_plot", now would be a great time to report it. (on- or off-list)
thanks
sven
4 months, 2 weeks
