Low-rank plus sparse estimation via Alternating Direction Method of Multipliers (ADMM)
An R package to fit the estimator suggested in Chandrasekaran et al. (https://projecteuclid.org/euclid.aos/1351602527). See the manual or the lrpsadmm-examples for examples and more details.
(Optional) Install the devtools package:
install.packages("devtools")Install the package:
library(devtools)
install_github("benjaminfrot/lrpsadmm")See the lrpsadmm-examples repository for step by step tutorials on how to use the package. The manual also contains examples.
Fit the Low-Rank plus Sparse (LRpS) estimator using the Alternating Direction Method of Multipliers (ADMM). This model learns an inverse covariance matrix which is the sum of a sparse matrix and a low-rank matrix as suggested by Chandrasekaran et al (2012) (DOI:10.1214/11-AOS949). The package supports robust estimation via an estimator of the correlation matrix based on Kendall rank correlations. It includes functions to compute a whole regularisation path and to select tuning parameters with cross-validation.
For questions and bug reporting, feel free to either send an email to benjamin dot frot @ gmail dot com or open an issue.