Skip to content

j-g-b/uncoupled

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

7 Commits
 
 
 
 
 
 
 
 

Repository files navigation

uncoupled

R package implementing minimum-Wasserstein isotonic regression estimator from Rigollet and Weed (2019)

Implementation details

To implement Rigollet and Weed's relaxed minimum-Wasserstein estimator, we use projected subgradient descent on the Wasserstein-2 objective. This amounts to a two-step procedure where evaluation of the objective at each iteration also yields subgradients that are combined to update the true measure weights. The deconvolved measure weights are then updated with a subgradient step and projected onto the probability simplex.

About

R package implementing minimum-Wasserstein isotonic regression estimator from Rigollet and Weed (2019)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages