Looking for the author of the Box_Cox_Value package, Pedro Isaac Chávez López
by Sven Schreiber
Hello everybody,
we have tried to contact Pedro Isaac Chávez López, who is the author of
the veteran contributed function package Box_Cox_Value. We used the
email addresses given in the help text of the package, but haven't
received an answer yet. But of course it's always possible that email
contacts have changed, or that our messages ended up in spam folders or
similar.
Is anybody here in touch with him? Perhaps you are even on this list, Pedro?
Any hints would be appreciated (on or off-list).
Thanks
Sven
2 months, 1 week
PDFdocs issue
by Clive
It could well be that a workaround has already been proposed, but:
$ sudo apt-get install texlive-latex-extra texlive-fonts-recommended
texlive-fonts-extra texlive-base texlive-latex-base texlive-lang-greek
texlive-xetex xterm git
$ git clone git://git.code.sf.net/p/gretl/git gretl-git
$ cd gretl-git
$ sudo apt-get build-dep gretl
$./configure --prefix=/usr/local --enable-quiet-build --enable-openmp
--enable-build-doc
$ make -j4 # all runs fine, but then ...
$ make pdfdocs
[...]
Package natbib Warning: Citation `koenker-zhao94' on page 74 undefined on
input
line 6503.
Package natbib Warning: Citation `koenker94' on page 74 undefined on input
line
6511.
Package natbib Warning: Citation `koenker-machado99' on page 74 undefined
on in
put line 6512.
[74]
! Missing $ inserted.
<inserted text>
$
l.6592 ...re separated by an underscore: eg snake_
case)
?
Thoughts?
Cheers, Clive
2 months, 1 week
Initial Release of the KNN Package
by Artur T.
Dear Gretl Community,
I am happy to announce the initial release of the KNN (K-Nearest
Neighbors) package for Gretl. This package introduces a comprehensive
suite of tools for performing both (multi-class) classification and
regression tasks using the KNN algorithm, directly within the Gretl
environment.
Key Features:
- **Flexible Model Fitting**: Fit your KNN model using a variety of data
splitting methods, including k-fold and leave-one-out cross-validation,
as well as recursive and rolling window techniques.
- **Standardization Support**: Automatically standardize features before
model fitting, with separate standardization for training and test sets
during cross-validation.
- **Comprehensive Scoring**: Evaluate model performance with a range of
metrics, including Precision, F-Score, and more, tailored to both
classification and regression tasks.
- **Optimal Neighbor Selection**: Utilize cross-validation to determine
the optimal number of neighbors for your model, enhancing prediction
accuracy.
- **GUI interface**: Call the interface via the menu "Model -> Robust
estimation --> KNN regression + classification"
Getting Started:
To install the KNN package, simply use the Gretl command interface or
the GUI package manager. For detailed documentation on using the
package, including function parameters and return values, please refer
to the included help file.
Happy Modeling,
Artur
2 months, 2 weeks
Store a bundles object
by Artur T.
Hi all,
I would like to store a bundles object (an array of bundles). However, I
only find the bwrite() function for storing a single bundle, but not a
bundles object. Am I right that it is currently not possible to store a
bundles object?
Thanks
Artur
2 months, 3 weeks
joint or combined test
by Olasehinde Timmy
Dear Professors,
I have been thinking recently about the possibility of combining F-test or
Chi-square test statistics from two different regression models to see
their joint significance. Particularly, let's say I have two ARDL models
with bound test values of 5.67 and 6.23 with k1 and k2 degrees of freedom.
I would like to know if it is possible to have a test or a way to combine
these two statistics to conclude their joint significance.
Please, advise me.
3 months
Version 0.7 of the fdensity Package
by Artur T.
Dear Gretl User-List,
I'm excited to announce the release of version 0.7 of the fdensity
package for Gretl. This package allows you to compute and visualize
conditional densities of a variable x for each value of a discrete
variable d, also known as the "factor."
The key changes in this new version are:
1. Support for list of series in x:
- If multiple series are passed to x, a grid of kernel density
estimation (KDE) plots will be created.
- This is currently only supported for the scripting mode. In GUI mode,
only a single series x can be selected.
2. New cumulative parameter:
- You can now compute the cumulative distribution function (CDF) instead
of the probability density function (PDF) by setting cumulative=TRUE.
3. New grid parameter:
- You can now control whether a grid is displayed on the plots by
setting grid=FALSE.
4. New logscale parameter:
- The y-axis can now be displayed on a logarithmic scale (base 10) by
setting logscale=TRUE.
5. GUI control parameters:
- The logscale and cumulative parameters are now exposed in the Gretl GUI.
6. Increased minimum version requirement:
- The minimum required Gretl version has been increased from 2021a to 2023c.
The help file incl. examples can be found here:
https://gretl.sourceforge.net/current_fnfiles/fdensity.gfn
Best,
Artur
3 months
Announcement: PairPlot Version 0.99 Release
by Artur T.
Dear gretl mailing list,
I'm pleased to announce the release of PairPlot version 0.99, which
includes the following new features:
- **Add new parameter title for setting an overall title**: Users can
now set a title for the overall plot.
- **Internal refactoring: Switch to gretl's built-in gridplot
apparatus**: This means that all the graphics file formats supported by
gretl can be used, and usage via `gretlcli` in "display" mode now works.
- **In case the user's specification calls for just a single plot, no
"multiplot" is created**: The package now handles single plots more
efficiently.
- **The graphical interface is enhanced, with more options**: Users will
find the updated package more user-friendly and customizable.
You can read the doc here:
https://gretl.sourceforge.net/current_fnfiles/unzipped/PairPlot.gfn
We encourage all gretl users to try out the new features and provide
feedback. If you have any questions or issues, please don't hesitate to
reach out to the gretl community.
Best,
Artur
3 months, 1 week
