Extending ERGM Functionality within statnet: Building Custom User Terms

Statnet Development Team
U of Washington, UC Irvine, Penn State, UCLA

LIMITED to 30 Seats

Exponential-family random graph models (ERGMs) represent a powerful and flexible class of models for the statistical analysis of networks. statnet is a software development project that includes a wide range of packages that support this modeling framework. The variables on the right hand side of an ERGM equation are different from the covariates in more traditional statistical models because they must be coded up by hand before they can be used in a model. statnet includes about 100 of the most commonly used terms in the ergm package; but if you want a specific term that is not included in the list, you would need to code it up yourself. This workshop teaches participants how to do this.

We provide a hands-on tutorial on the template statnet package “ergm.userterms” that can be modified to create user-specific ERGM terms. It is designed to make this process as straightforward as possible. We explain some of the internal workings of statnet that will help users develop their own network analysis capabilities. The workshop will teach through examples in a tutorial paper and demonstrate each step in the practical process. Participants work in small groups to code up their own ergm.userterms during the last part of the workshop.

Prerequisites: This workshop will assume familiarity with R, and the ergm package in statnet. The “Exponential-family Random Graph (ERG or p*). Modeling with statnet” workshop is recommended as preparation.

statnet is a collection of packages for the R statistical computing environment that supports the representation, manipulation, visualization, modeling, simulation, and analysis of network data. statnet packages are contributed by a team of volunteer developers, and are made freely available under the GNU Public License. statnet packages can be used with any computing platform that supports R (including Windows, Linux, and Mac), and they support statistical analysis of large networks, longitudinal network dynamics, and missing data.