New Gretl package: pcaTools
by Artur T.
Hello Gretl community,
I would like to announce a new package for conducting Principal
Component Analysis (PCA). This package not only allows you to perform
PCA but also includes some additional features that can be quite useful
for your analysis.
- Scree Plot: The package supports a scree plot, a graphical tool used
in PCA to determine the number of principal components to retain.
- Biplot: The package includes support for a biplot, a type of graph
used in PCA to visualize the relationships between the data. A biplot
allows you to see both the scores of the samples and the loadings of the
variables, providing a comprehensive view of the data.
Please note that this package is not officially available on the Gretl
package server. However, you can find the Gretl package file on GitHub
at this link:
https://github.com/atecon/pcaTools/blob/main/src/pcaTools.gfn
You can find the documentation for this package here:
https://github.com/atecon/pcaTools/blob/main/src/pcaTools_help.md
A sample script demonstrating how to use the package is also available:
https://github.com/atecon/pcaTools/blob/main/src/pcaTools_sample.inp
Gretl has built-in commands (`pca`) for estimating the principal
components. This new package provides additional tools for PCA but does
not replace the built-in functionality.
There is also the `staticfactor` package available in Gretl, which
supports Principal Component analysis using advanced methods but it
(currently) has not support for creating bi-plots.
I hope you find this new package useful for your PCA needs.
Best,
Artur
8 months
Digits - rounding problem
by Meridiana GeoTopo
Hi everyone. Wish you all a Happy 2024, first of all.
I wonder if there is a way to make gretl work with more decimal digits. I
think I am having a rounding error problem.
I was getting divergent $coeff so I ran against other calculations OLS vs
matrix solutions in gretl - against some book examples
Example::
GRETL output:
? mX
mX (4 x 2) [t1 = 101, t2 = 104]
0,83732 -1,2049
1,4338 -0,22646
-0,58424 1,5486
-1,7969 0,57819
? f
f
-3,92 5,43 -5,94 13,52
? B == mX' * mX
B (2 x 2)
6,3271 -3,2773
-3,2773 4,2355
? T == B' * f
T (2 x 1)
-16,321
2,1119
? B^-1
0,26377 0,20409
0,20409 0,39402
? result
*result (2 x 1) == B^-1 * T*
* -3,8738 *
* -2,4988 *
*Now against the "benchmark' :*
*B =*
6.327065 -3.277288
-3.277288 4.235533
*T =*
-16.323890
2.114714
*B^-1 =*
0.263767 0.204092
0.204092 0.394016
*result =*
*-3.8741*
*-2.4983*
8 months, 1 week