The R
package BDgraph provides statistical tools for Bayesian structure learning for undirected graphical models with 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. (2021), Mohammadi et al. (2017), and Dobra and Mohammadi (2018). Besides, the package contains several functions for simulation and visualization, as well as several multivariate datasets taken from the literature.
Install BDgraph using
install.packages( "BDgraph" )
First, we load BDgraph package
library( BDgraph )
Here are two simple examples to show how to use the functionality of the package.
Here is a simple example to see the performance of the package for the Gaussian graphical models. First, by using the function bdgraph.sim()
, we simulate 200 observations (n = 200) from a multivariate Gaussian distribution with 15 variables (p = 15) and “scale-free” graph structure, as follows
set.seed( 20 )
= bdgraph.sim( n = 200, p = 15, graph = "scale-free", vis = TRUE ) data.sim
Since the generated data are Gaussian, we run the bdgraph()
function by choosing method = "ggm"
, as follows
= bdgraph( data = data.sim, method = "ggm", iter = 5000, verbose = FALSE ) bdgraph.obj
To report confusion matrix with cutoff point 0.5:
conf.mat( actual = data.sim, pred = bdgraph.obj, cutoff = 0.5 )
Actual0 1
Prediction 0 91 4
1 0 10
conf.mat.plot( actual = data.sim, pred = bdgraph.obj, cutoff = 0.5 )
To compare the result with the true graph
compare( data.sim, bdgraph.obj, main = c( "Target", "BDgraph" ), vis = TRUE )
Target BDgraph
True Positive 10 10.000
True Negative 95 91.000
False Positive 0 4.000
False Negative 0 0.000
F1-score 1 0.833
Specificity 1 0.958
Sensitivity 1 1.000
MCC 1 0.827
Now, as an alternative, we run the bdgraph.mpl()
function which is based on the GGMs and marginal pseudo-likelihood, as follows
= bdgraph.mpl( data = data.sim, method = "ggm", iter = 5000, verbose = FALSE )
bdgraph.mpl.obj
conf.mat( actual = data.sim, pred = bdgraph.mpl.obj )
Actual0 1
Prediction 0 91 4
1 0 10
conf.mat.plot( actual = data.sim, pred = bdgraph.mpl.obj )
We could compare the results of both algorithms with the true graph as follows
compare( list( bdgraph.obj, bdgraph.mpl.obj ), data.sim,
main = c( "Target", "BDgraph", "BDgraph.mpl" ), vis = TRUE )
Target BDgraph BDgraph.mpl
True Positive 14 10.000 10.000
True Negative 91 91.000 91.000
False Positive 0 0.000 0.000
False Negative 0 4.000 4.000
F1-score 1 0.833 0.833
Specificity 1 1.000 1.000
Sensitivity 1 0.714 0.714
MCC 1 0.827 0.827
To see the performance of the BDMCMC algorithm we could plot the ROC curve as follows
plotroc( list( bdgraph.obj, bdgraph.mpl.obj ), data.sim, cut = 200,
labels = c( "BDgraph", "BDgraph.mpl" ), color = c( "blue", "red" ) )
Here is a simple example to see the performance of the package for the mixed data using Gaussian copula graphical models. First, by using the function bdgraph.sim()
, we simulate 300 observations (n = 300) from mixed data (type = "mixed"
) with 10 variables (p = 10) and “random” graph structure, as follows
set.seed( 2 )
= bdgraph.sim( n = 300, p = 10, type = "mixed", graph = "random", vis = TRUE ) data.sim
Since the generated data are mixed data, we are using run the bdgraph()
function by choosing method = "gcgm"
, as follows:
= bdgraph( data = data.sim, method = "gcgm", iter = 5000, verbose = FALSE ) bdgraph.obj
To compare the result with the true graph, we could run
compare( bdgraph.obj, data.sim, main = c( "Target", "BDgraph" ), vis = TRUE )
Target BDgraph
True Positive 12 10.000
True Negative 33 32.000
False Positive 0 1.000
False Negative 0 2.000
F1-score 1 0.870
Specificity 1 0.970
Sensitivity 1 0.833
MCC 1 0.826
plotroc( bdgraph.obj, data.sim, labels = "BDgraph", color = "blue" )
For more examples see Mohammadi and Wit (2019).