- Notifications
You must be signed in to change notification settings - Fork0
Comparisons between best subset selection and other popular estimators for sparse regression
License
guhjy/best-subset
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
This project contains an implementation of best subset selection inregression, based on a mixed integer quadratic program formulation ofthe subset selection problem and the Gurobi mixed integer programoptimizer. It also contains tools for running simulations comparingbest subset selection to other common sparse regression estimatorssuch as the lasso and forward stepwise selection.
The mixed integer programming formulation of subset selection andsimulation setup is based on the paper: Bertsimas, King, Mazumder(2016), "Best subset selection via a modern optimization lens".
For our discussion paper on best subset selection, forward stepwise,and the lasso, see:https://arxiv.org/abs/1707.08692.
To install the bestsubset R package directly from github, run thefollowing in R:
library(devtools)install_github(repo="ryantibs/best-subset", subdir="bestsubset")