This software, written in the R language,
estimates a sparse inverse covariance matrix using a lasso (L1) penalty.
It can be used for estimating a sparse undirected graph.
Version 1.3 fixed some errors in the Meinhausen-Buhlmann approximation, and includes a new function for computing the solution along a path of regularization parameters.
Based on the paper
Jerome Friedman, Trevor Hastie and Robert Tibshirani Sparse inverse covariance estimation with the graphical lasso. Biostatistics, December 12, 2007.
Linux, Windows and MAC versions available at CRAN