col_gen_estimator - A template for scikit-learn compatible column generation based estimators contributions ============================================================
col_gen_estimator is a template project for scikit-learn compatible column generation based estimators.
This project is built using the sklearn template.
It aids development of estimators that can be used in scikit-learn pipelines and (hyper)parameter search, while facilitating testing (including some API compliance), documentation, open source development, packaging, and continuous integration.
Following example extensions of a column generation based binary classifiers are included. - Boolean Decision Rule Generation by S. Dash et. al. 2018 - Decision Tree classifiers by Firat et. al. 2020 (modified)
The developer needs to extend the master and subproblem classes and implement the required methods. The coumn generation part is taken care of by the template fit method.
The Decision Tree Classifier experiments can be launched from the example directory. See the README file in the examples directory for details.
To reproduce the results submitted in the Decision tree paper, use the branch 'dtreedev'.
Thank you for cleanly contributing to the estimator template!