ergm package is capable of fitting a wide range of
exponential random network models, in which
the probability of a given network, $y$, on a set of nodes is
$\exp(\theta{\cdot}g(y))/c(\theta)$, where
$g(y)$ is a vector of network statistics,
$\theta$ is a parameter vector of the same length and $c(\theta)$ is the
normalizing constant for the distribution.
The ergm function fits these models when they are expressed
via an Rformula object, of the form
y ~ ,
where y is a network object or a matrix that can be
coerced to a network object.
To create a
network object in R, use the network() function,
then add nodal attributes to it using the %v%
operator if necessary. The ergm package contains a wide range of terms.
For the details on the possible
ergm-terms.
This package can be modified by users to add user-defined terms to ergm
models.
The terms can be used throughout the ergm package
and behave identically to the supplied terms.
ergmuserterms package is available from the statnet
website ( The code contains some simple examples and templates.
These include:
m2starkstar(2). This
option adds one statistic to the model, equal to the
number of mixed-2-stars in the network, defined as a
pair of edges ${(i{\rightarrow}j), (j{\rightarrow}k)}.$
}
testmeedges is isomorphic to kstar(1); for
directed networks, edges is isomorphic to both ostar(1)
and istar(1). }
In the implementation of ergm, the model is initialized in R,
then all the model information is passed to a C program that generates
the sample of graph statistics using MCMC. This sample is then returned
to R, which then approximates the MLE.
library(ergmuserterms)
data(sampson)
monk.fit <- ergm(samplike~m2star)
summary(monk.fit)
monk.fit <- ergm(samplike ~ m2star + testme)
summary(monk.fit)Run the code above in your browser using DataLab