This is a Python adaptation of the standalone Matlab function that calculates the Lasso Trade-off Diagram introduced in the paper of Su et al., 2017, "False Discoveries Occur Early on the Lasso Path" (http://arxiv.org/abs/1511.01957).
Source for the original Matlab function: https://github.com/wjsu/fdrlasso
The paper_implementation Jupyter notebook has been developed with the aim of reproducing the authors' results, but also to extend them to the framework of the positive Lasso.
This work was carried out as part of a project for the Master course Theoretical guidelines for high dimensional data analysis, taught by Christophe Giraud, professor at the Université Paris-Sud.
In no way do the authors of this notebook claim any theoretical or empirical discovery of the results presented.
Authors: Clotilde Miura and Cédric Allain