Dear Maintainers,
I work on sparse PLS from now many years (doc+postDoc in INRIA and INSERM, see
[ https://hadrienlorenzo.netlify.app/ | https://hadrienlorenzo.netlify.app ]
for light view) and published about applications. Main problems are about
dealing with missing values in the multi-output and degenerate n<<p contexts
for multi-block structures.
I wrote packages
* in R : [ https://cran.r-project.org/web/packages/ddsPLS/index.html |
https://cran.r-project.org/web/packages/ddsPLS/index.html ] ,
* and in Python : [ https://pypi.org/project/py-ddspls/ |
https://pypi.org/project/py-ddspls/ ] .
In the both objectives of offering sparse PLS opportunities to the Python
community and to improve my Python skills, I would like to propose you a python
version of the algorithm for which specificities are the following.
* Modification of the PLS2 algorithm.
* Soft-thresholding of the empirical covariance matrices.
* Automatic-running of the number of component and the the sparsity
parameter through bootstrap sampling.
* Sparsity both in X and in Y with a single parameter.
This algorithm has been developed with Jérôme Sarraco (Pr INRIA) and Rodolphe
Thiébaut (PUPH Inserm) and recently with Olivier Cloarec (Sartorius, Research
Group for Chemometrics, Institute of Chemistry, Umeå University, Umeå, S-901
87, Suède)
Would you be interested in this sparse version of the PLS algorithm ? I am more
than eager to discuss about this project with you, so do not hesitate to
contact me.
Bests,
Hadrien Lorenzo
+33 6 49 09 55 78
_______________________________________________
scikit-learn mailing list
scikit-learn@python.org
https://mail.python.org/mailman/listinfo/scikit-learn
