huge: High-dimensional Undirected Graph Estimation

The package "huge" provides a general framework for high-dimensional undirected graph estimation. It integrates data preprocessing, neighborhood screening, graph estimation, and model selection techniques into a pipeline. In preprocessing stage, the nonparanormal(npn) transformation is applied to help relax the normality assumption. In the graph estimation stage, the graph structure is estimated by Meinshausen-Buhlmann graph estimation or the graphical lasso, and both methods can be further accelerated by the lossy screening rule preselecting the neighborhood of each variable by correlation thresholding. We target on high-dimensional data analysis usually d >> n, and the computation is memory-optimized using the sparse matrix output. We also provide a computationally efficient approach, correlation thresholding graph estimation. Three regularization/thresholding parameter selection methods are included in this package: (1)stability approach for regularization selection (2) rotation information criterion (3) extended Bayesian information criterion which is only available for the graphical lasso.

Version: 1.2.6
Depends: R (≥ 3.0.0), Matrix, lattice, igraph, MASS
Published: 2014-02-28
Author: Tuo Zhao, Han Liu, Kathryn Roeder, John Lafferty, Larry Wasserman
Maintainer: Tuo Zhao <tzhao5 at>
License: GPL-2
NeedsCompilation: yes
CRAN checks: huge results


Reference manual: huge.pdf
Vignettes: vignette
Package source: huge_1.2.6.tar.gz
Windows binaries: r-devel:, r-release:, r-oldrel:
OS X Snow Leopard binaries: r-release: huge_1.2.6.tgz, r-oldrel: huge_1.2.6.tgz
OS X Mavericks binaries: r-release: huge_1.2.6.tgz
Old sources: huge archive

Reverse dependencies:

Reverse imports: qgraph
Reverse suggests: LICORS, sand, smart, stm