An introduction to statnet: an R-based program for the statistical analysis and simulation of social networks

To date, the use of stochastic network models for networks has been limited by three interrelated factors: the complexity of realistic models, the lack of simulation tools for inference and validation, and a poor understanding of the inferential properties of nontrivial models. This seminar provides an overview of statnet, a statistical package for the visualization, analysis and simulation of social network data. The modeling capabilities of statnet include the class of exponential random graph models and latent space and cluster models. These models recognize the complex dependencies within relational data structures, and provide a very flexible framework for representing them. Examples include degree distributions and stars, attribute-based mixing patterns, triadic patterns that lead to clustering, shared partner distributions, and other systematic network configurations. statnet has a coherent and flexible user interface and can handle relatively large networks, and it has very efficient algorithms for data manipulation and analysis. The package provides tools for both model estimation and model-based network simulation, with visualization, tools for inference and validation, and goodness of fit diagnostics. The package is written for the R statistical computing environment, so it runs on any computing platform that supports R (Windows, Unix/Linux, Mac). Detailed information on the package is available at http://csde.washington.edu/statnet