This version has example config with self-edges (loops) and a bug fix

Examples of config files with loops (self-edges).
Correction for max num edges if both loops allowed and reciprocity not.

This version allows self-edges (loops)

This version allows self-edges (loops) if the allowLoops configuration parameter is set to True.
Also fixes a bug with IFD sampler arc correction and IFD/TNT sampler acceptance probability if forbidReciprocity = True.
New scripts to compute Malahanobis distance of observed from simulated networks.

Include TNT and IFD sampler conditional estimation, simulation, and goodndess-of-fit for the "citation ERGM" (cERGM) of Schmid, C. S., Chen, T. H. Y., & Desmarais, B. A. (2021)

This release includes implementation of conditional estimation in the TNT and IFD samplers, as well as simulation and goodness-of-fit, for the "citation ERGM" (cERGM) model described by:

Schmid, C. S., Chen, T. H. Y., & Desmarais, B. A. (2021). Generative Dynamics of Supreme Court Citations: Analysis with a New Statistical Network Model. arXiv preprint arXiv:2101.07197.

The original statnet (http://statnet.org/) R implementation of cERGM is available from https://github.com/schmid86/cERGM/.

Implement the "citation ERGM" (cERGM) of Schmid, C. S., Chen, T. H. Y., & Desmarais, B. A. (2021)

This release includes conditional estimation and statistics to implement the "citation ERGM" (cERGM) model described by:

Schmid, C. S., Chen, T. H. Y., & Desmarais, B. A. (2021). Generative Dynamics of Supreme Court Citations: Analysis with a New Statistical Network Model. arXiv preprint arXiv:2101.07197.

The original statnet (http://statnet.org/) R implementation of cERGM is available from https://github.com/schmid86/cERGM/.

Ability to set lambda values on alternating parameters

Ability to set lambda values on alternating parameters e.g. AltKTriangles(2.5) to set lambda = 2.5. The default remains 2.0.
Also improvements to goodness-of-fit plot including -s option to put triad census in separate .eps file for better visibility.

Update R scripts to work with R package mcse 1.4-1

use size='sqroot' on mcse.multi for same behaviour as old version

(prior to 1.4-1 when it did not have r parameter and size default was sqrt)

Minor bug fixes

Fix bug with initial value of observed stats that are not zero (i.e. Isolates) for computeObservedStats.
Fix crowding of x axis labels on simulation/GoF plots for plotEstimNetDirecteSimFit.R script
Merge in faster arclist usage changes branch (makes little difference in practice).

This release is more scalable in memory usage

This release now allows (by default) change statistic computation without using lookup tables (either hash tables or arrays), which is much more scalable in terms of memory usage. This can also be more efficient in some cases even when large memory usage was not required.

This release includes ERGM simulation and goodness-of-fit scripts

Includes program to simulate ERGM networks, and scripts for goodness-of-fit tests. Also the TNT (Tie-No-Tie) sampler, Euclidean distance dyadic covariate, and change statistic for Jaccard similarity between sets of categorical attributes on nodes.

Second release includes parameters for citation networks

This release includes the DiffSign parameter for citation temporal direction in citation networks, and an option to forbid reciprocity in the simulation process, also useful for citation networks. It also includes standard error estimation using the mcmcse R package.