BDgraph

The R package BDgraph provides statistical tools for Bayesian structure learning in undirected graphical models for continuous, count, binary, and mixed data. The package is implemented the recent improvements in the Bayesian graphical models literature, including Mohammadi and Wit (2015), Mohammadi et al. (2017), Dobra and Mohammadi (2018), and Letac et al. (2018). Besides, the package contains several functions for simulation and visualization, as well as several multivariate datasets taken from the literature. To speed up the computations, the computationally intensive tasks of the package are implemented in C++ in parallel using OpenMP.

Installation

You can install the latest version from CRAN using:

install.packages( "BDgraph" )
require( "BDgraph" )

Example 1: Gaussian Graphical Models

Here is a simple example to see the preformance of the package for the Gaussian graphical models. Frist, by using the function bdgraph.sim to simulate 100 observations (n = 100) from a multivariate Gaussian distribution with 8 variables (p = 8) and “scale-free” graph structure, as follows:

data.sim = bdgraph.sim( n = 100, p = 8, graph = "scale-free", vis = TRUE )
round( head( data.sim $ data, 4 ), 2 )

Since the generated data are Gaussian, we run the bdgraph function by choosing method = "ggm", as follows:

bdgraph.obj <- bdgraph( data = data.sim, method = "ggm", iter = 5000 )

summary( ssgraph.obj )

To compare the result with the true graph

compare( data.sim, bdgraph.obj, main = c( "Target", "BDgraph" ), vis = TRUE )

Now, as an alternative, we run the bdgraph.mpl function which is based on the GGMs and marginal pseudo-likelihood, as follows:

bdgraph.obj_mpl <- bdgraph.mpl( data = data.sim, method = "ggm", iter = 5000 )

summary( bdgraph.obj_mpl )

We could compare the results of both algorithms with the true graph as follows:

compare( data.sim, bdgraph.obj, bdgraph.obj_mpl, 
         main = c( "Target", "BDgraph", "BDgraph_mpl" ), vis = TRUE )

Example 2: Gaussian Copula Graphical Models

Here is a simple example to see the preformance of the package for the mixed data using Gaussian copula graphical models. Frist, by using the function bdgraph.sim to simulate 100 observations (n = 100) from mixed data (type = "mixed") with 7 variables (p = 7) and “random” graph structure, as follows:

data.sim = bdgraph.sim( n = 100, p = 7, type = "mixed", graph = "random", vis = TRUE )
round( head( data.sim $ data, 4 ), 2 )

Since the generated data are mixed data, we are using run the bdgraph function by choosing method = "gcgm", as follows:

bdgraph.obj <- bdgraph( data = data.sim, method = "gcgm", iter = 5000 )

summary( ssgraph.obj )

To compare the result with the true graph, we could run

compare( data.sim, bdgraph.obj, main = c( "Target", "BDgraph" ), vis = TRUE )

For more examples see Mohammadi and Wit (2019).