ADMM
Algorithms using Alternating Direction Method of Multipliers
Provides algorithms to solve popular optimization problems in statistics such as regression or denoising based on Alternating Direction Method of Multipliers (ADMM).
See Boyd et al (2010)
ADMM
We provide implementation for a class of problems that use alternating direction method of multipliers (ADMM)-type algorithms.
Installation
You can install the released version of ADMM from CRAN with:
install.packages("ADMM")
And the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("kyoustat/ADMM")
Available Functions
Currently, we support following classes of problems and functions. For
more details, please see help pages of each function using help()
function in your R session.
| Function | Description |
|---|---|
admm.bp |
Basis Pursuit |
admm.enet |
Elastic Net Regularization |
admm.genlasso |
Generalized LASSO |
admm.lad |
Least Absolute Deviations |
admm.lasso |
Least Absolute Shrinkage and Selection Operator |
admm.rpca |
Robust Principal Component Analysis |
admm.sdp |
Semidefinite Programming |
admm.spca |
Sparse Principal Component Analysis |
admm.tv |
Total Variation Minimization |