glassomix: High dimensional Mixture Graph Models selection

The package glassomix provides a general framework for network recovering through a model-based soft clustering. It provides functions for parameter estimation via the EM algorithm for Gaussian graphical mixture models in high dimensional setting. The main function is “glasso.mix” upon which a model selection is performed. The package estimates the optimum number of mixture components, K and the tuning parameter, lambda, based on the Extended Bayesian Information Criteria (EBIC) via “select.gm” function. The graphical structural of the K-networks are also plotted through the function “gm.plot”

Version: 1.2
Depends: R (≥ 3.0.0), mvtnorm, glasso, huge
Published: 2013-11-05
Author: Anani Lotsi and Ernst Wit
Maintainer: Anani Lotsi <a.lotsi at rug.nl>
License: GPL (≥ 3)
URL: http://www.math.rug.nl/stat/Main/Research
NeedsCompilation: no
CRAN checks: glassomix results

Downloads:

Reference manual: glassomix.pdf
Package source: glassomix_1.2.tar.gz
MacOS X binary: glassomix_1.2.tgz
Windows binary: glassomix_1.2.zip
Old sources: glassomix archive