Wikipedia defines a graphical model as a graph that represents independencies among random variables by a graph in which each node is a random variable, and the missing edges between the nodes represent conditional independencies .

This task view is a collection of packages intended to supply R code to deal with graphical models. The description of each package is accompanied by the statement [vignette: y/n]. The aim is to give users indications of to which extent a package is documented beyond the bare standard documentation.

The packages can be roughly structured into the following topics (although several of them have functionalities which go across these categories):

Representation, manipulation and display of graphs

• diagram [vignette: y] Functions for visualising simple graphs (networks), plotting flow diagrams
• dynamicGraph [vignette: n] Provides an advanced graphical user interface for display and interaction with graphs (based on Tcl/Tk).
• giRaph [vignette: n] Supply classes and methods to represent and manipulate graphs
• graph [vignette: y] A package to handle graph data structures.
• gRbase [vignette: y] An alternative interface to graphs. Uses the graph and RBGL and implements additional graph operations.
• igraph [vignette: n] Routines for simple graphs, network analysis.
• mathgraph [vignette: n] Implements matrix representations of graphs and provide a plot function.
• network [vignette: n; demo: n] Tools to create and modify network objects. The network class can represent a range of relational data types, and supports arbitrary vertex/edge/graph attributes.
• RBGL [vignette: y] Interface to boost C++ graph library (based on graph objects from the graph package).

In addition, there is the Rgraphviz package on Bioconductor that provides plotting capabilities for R graph objects (from the graph package) [vignette: y].

Classical models - General purpose packages

• ggm [vignette: n] Fitting graphical Gaussian models.
• gRbase [vignette: y] Sets up a meta data representation which is used by several other packages (among these are mimR and gRain). gRbase suggests a way to implement models and fitting engines - illustrated by hierarchical log-linear models.
• mimR [vignette: n] General package for model selection in contingency tabels, graphical Gaussian models and mixed interaction models. The package provides an interface to the MIM program and only runs on Windows.

Miscellaneous: Model search, specialized types of models etc.

• SIN [vignette: n] This package provides routines to perform SIN model selection as described in Drton and Perlman (2004). The selected models are represented in the format of the 'ggm' package, which allows in particular parameter estimation in the selected model.
• pcalg [vignette: y] Standard and robust estimation of the skeleton (ugraph) and the equivalence class of a Directed Acyclic Graph (DAG) via the PC-Algorithm. The equivalence class is represented by its (unique) Completed Partially Directed Acyclic Graph (CPDAG).
• qp [vignette: n] This package is deprecated and it is now only a stub for the newer version called qpgraph available through the Bioconductor project. The q-order partial correlation graph search algorithm, q-partial, or qp, algorithm for short, is a robust procedure for structure learning of undirected Gaussian graphical Markov models from "small n, large p" data, that is, multivariate normal data coming from a number of random variables p larger than the number of multidimensional data points n as in the case of, e.g., microarray data.
• gRc [vignette: n] Inference in graphical Gaussian models with edge and vertex symmetries (Graphical Gaussian models with colours).
• GeneNet [vignette: n] Modeling and Inferring Gene Networks. GeneNet is a package for analyzing gene expression (time series) data with focus on the inference of gene networks.
• parcor [vignette: n] The package estimates the matrix of partial correlations based on different regularized regression methods: lasso, adaptive lasso, PLS, and Ridge Regression. In addition, the package provides model selection for lasso, adaptive lasso and Ridge regression based on cross-validation.
• catnet [vignette: y] A package that handles discrete Bayesian network models and provides inference using the frequentist approach

Bayesian Networks/Probabilistic expert systems

• bnlearn [vignette: n] Bayesian network structure learning via constraint-based (also known as 'conditional independence') and score-based algorithms. This package implements the Grow-Shrink (GS) algorithm, the Incremental Association (IAMB) algorithm, the Interleaved-IAMB (Inter-IAMB) algorithm, the Fast-IAMB (Fast-IAMB) algorithm, the Max-Min Parents and Children (MMPC) algorithm and the Hill-Climbing (HC) greedy search algorithm for both discrete and Gaussian networks, along with many score functions and conditional independence tests. Some utility functions (model comparison and manipulation, random data generation, arc orientation testing) are also included.
• deal [vignette: n] Learning Bayesian networks with mixed (discrete and continuous) variables.
• gRain [vignette: y] A package for probability propagation in graphical independence networks, also known as probabilistic expert systems (which includes Bayesian networks as a special case).

BUGS models

• bayesmix [vignette: n; demo: n] Bayesian mixture models of univariate Gaussian distributions using JAGS
• boa [vignette: n; demo: n] Bayesian Output Analysis Program (BOA) for MCMC.
• BRugs [vignette: n; demo: n] An R / S-PLUS package containing OpenBUGS and its R / S-PLUS interface BRugs. Notice: currently the only supported OS is Windows, we expect to support Linux in future releases.
• coda [vignette: n; demo: n] Output analysis and diagnostics for Markov Chain Monte Carlo simulations.
• ergm [vignette: n; demo: n] An integrated set of tools to analyze and simulate networks based on exponential-family random graph models (ERGM). "ergm" is a part of the statnet suite of packages for network analysis.
• G1DBN [vignette: n; demo: n] G1DBN performs DBN inference using 1st order conditional dependencies.
• rbugs [vignette: n; demo: n] Fusing R and OpenBugs
• R2WinBUGS [vignette: y; demo: n] Running WinBUGS and OpenBUGS from R / S-PLUS