ggm: The package 'ggm': summary information

In ggm: Graphical Markov Models with Mixed Graphs

Description

This package provides functions for defining, manipulating and fitting graphical Markov models with mixed graphs. It is intended as a contribution to the gR-project described by Lauritzen (2002).

For a tutorial illustrating the new functions in the package 'ggm' that deal with ancestral, summary and ribbonless graphs see Sadeghi and Marchetti (2012) in the references.

Functions

The main functions can be classified as follows.

• Functions for defining graphs (undirected, directed acyclic, ancestral and summary graphs): UG, DAG, makeMG, grMAT;

• Functions for doing graph operations (parents, boundary, cliques, connected components, fundamental cycles, d-separation, m-separation): pa, bd, cliques, conComp, fundCycles;

• Functions for testing independence statements and generating maximal graphs from non-maximal graphs: dSep, msep, Max;

• Function for finding covariance and concentration graphs induced by marginalization and conditioning: inducedCovGraph, inducedConGraph;

• Functions for finding multivariate regression graphs and chain graphs induced by marginalization and conditioning: inducedRegGraph, inducedChainGraph, inducedDAG;

• Functions for finding stable mixed graphs (ancestral, summary and ribbonless) after marginalization and conditioning: AG, SG, RG;

• Functions for fitting by ML Gaussian DAGs, concentration graphs, covariance graphs and ancestral graphs: fitDag, fitConGraph, fitCovGraph, fitAncestralGraph;

• Functions for testing several conditional independences:shipley.test;

• Functions for checking global identification of DAG Gaussian models with one latent variable (Stanghellini-Vicard's condition for concentration graphs, new sufficient conditions for DAGs): isGident, checkIdent;

• Functions for fitting Gaussian DAG models with one latent variable: fitDagLatent;

• Functions for testing Markov equivalences and generating Markov equivalent graphs of specific types: MarkEqRcg, MarkEqMag, RepMarDAG, RepMarUG, RepMarBG.

Authors

Giovanni M. Marchetti, Dipartimento di Statistica, Informatica, Applicazioni 'G. Parenti'. University of Florence, Italy

Mathias Drton, Department of Statistics, University of Washington, USA

Kayvan Sadeghi, Department of Statistics, Carnegie Mellon University, USA

Acknowledgements

Many thanks to Fulvia Pennoni for testing some of the functions, to Elena Stanghellini for discussion and examples and to Claus Dethlefsen and Jens Henrik Badsberg for suggestions and corrections. The function fitConGraph was corrected by Ilaria Carobbi. Helpful discussions with Steffen Lauritzen and Nanny Wermuth, are gratefully acknowledged. Thanks also to Michael Perlman, Thomas Richardson and David Edwards.

Giovanni Marchetti has been supported by MIUR, Italy, under grant scheme PRIN 2002, and Mathias Drton has been supported by NSF grant DMS-9972008 and University of Washington RRF grant 65-3010.

References

Lauritzen, S. L. (2002). gRaphical Models in R. R News, 3(2)39.

Sadeghi, K. and Marchetti, G.M. (2012). Graphical Markov models with mixed graphs in R. The R Journal, 4(2):65-73. https://journal.r-project.org/archive/2012/RJ-2012-015/RJ-2012-015.pdf

ggm documentation built on March 26, 2020, 7:49 p.m.