From 80b358128548567b7b7a3c19f261b995d22bfed4 Mon Sep 17 00:00:00 2001 From: Martina Morris Date: Mon, 27 Nov 2023 15:32:08 -0800 Subject: [PATCH] Update README.md Added a more user friendly description of package utility, and links to the tergm workshop, 4.0 crosswalk vignette and citation. This updates the README to the new format. --- README.md | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index 59e520a4..a11dfa4b 100644 --- a/README.md +++ b/README.md @@ -5,7 +5,32 @@ [![Coverage status](https://codecov.io/gh/statnet/tergm/branch/master/graph/badge.svg)](https://codecov.io/github/statnet/tergm?branch=master) [![R build status](https://github.com/statnet/tergm/workflows/R-CMD-check/badge.svg)](https://github.com/statnet/tergm/actions) -An integrated set of extensions to the 'ergm' package to analyze and simulate network evolution based on exponential-family random graph models (ERGM). 'tergm' is a part of the 'statnet' suite of packages for network analysis. See Krivitsky and Handcock (2014) and Carnegie, Krivitsky, Hunter, and Goodreau (2015) . +An integrated set of extensions to the 'ergm' package to analyze and simulate dynamic temporal networks based on exponential-family random graph models (ERGM). 'tergm' is a part of the Statnet suite of packages for network analysis. + +TERGMs are a broad, flexible class of models for representing temporal edge dependence structure and dynamics. They extend the generalized linear model framework by relaxing the assumption that observations (dyads) are independent. Dependence is explicitly represented by model terms (e.g. mutuality, degree distributions, types of triads, etc) that can be fit along with dyad-independent terms (e.g., homophily). Statistical inference supports significance testing for the model and parameter estimates. `tergm` can be used for both estimation from and simulation of dynamic networks -- both workflows rely on Statnet's central MCMC algorithm, which can be tuned with a detailed set of control parameters. + +Temporal network data come in different forms, continuous and discrete, completely observed and sampled. The tergm package is designed to work with binary or continuously valued edges for: + +* Network panel data +* One cross-sectional network with continuous edge duration information +* One cross-sectional, egocentrically sampled network with continuous edge duration + +Tools are included for implementing a standard statistical workflow: + +* Exploratory data analysis -- via the related Statnet packages [`tsna`](https://github.com/statnet/tsna/) and [`ndtv`](https://github.com/statnet/ndtv/) +* Model estimation (data-type specific) +* Model diagnostics for assessing convergence and "Goodness of Fit" +* Simulating dynamic networks (from fitted or hypothetical models) + +## Docs and examples + +A brief introduction to the TERGM statistical framework, along with many examples of how `tergm` tools might be used in a network data analysis workflow can be found in our self-guided workshop materials: [Temporal Exponential Random Graph Models (TERGMs) for dynamic networks](https://statnet.org/workshop-tergm/). + +Note that with the release of `tergm` 4.0, the API has been modified to allow for the specification of a much wider class of models. A crosswalk of the differences can be found in the [`tergm` Conversion vignette](https://cran.r-project.org/web/packages/tergm/vignettes/tergm4_conversion.html) + +## Citation + +Citation information (with bibtex entries) can be found [here](https://cran.r-project.org/web/packages/tergm/citation.html) ## Public and Private repositories