Stochastic Models

This course aims to provide a practical introduction into cross-sectional ERGM (p* models) and longitudinal RSIENA models with a focus on hands-on applications of programs such as PNET and RSIENA and the interpretation of the results.

The course starts with a discussion of statistical inference for complete network analysis and some simple statistical tests are run in class. We then discuss more complex models, with a specific focus on ERGM (p* models) for cross-sectional social network data and RSIENA for longitudinal social network data.

Software

• • PNet (http://www.sna.unimelb.edu.au/pnet/pnet.html). Both PNet.jar and pnet.dll are needed.

  • R for SIENA and the stochastic tests (http://cran.r-project.org/)

    • Download R for Windows or other is needed. SPecific packages (RSIENA and sna) will best be downloaded from internet at the start of the course (to ensure everyone has the same version!). If needed they can be put on a usb-stick.

Preliminary schedule

• • Day 1/Morning: Random graphs and statistical tests

  • Day 1/Afternoon: Intro into ERG models

  • Day 2/Morning: Running ERG (p*) models with PNet

  • Day 2/Afternoon: Focus on interpretation of parameters for different models

  • Day 3/Morning: More interpretation of ERGM and more advanced topics

  • Day 3/Afternoon: Siena models with RSIENA

  • Day 4/Morning: Interpretation of parameters from RSIENA

  • Day 4/Afternoon: Interpretation of parameters from RSIENA and more advanced topics

Readings

Robins, G., P. Pattison, Y. Kalish, and D. Lusher (2007). On exponential random graph models for cross-sectional analysis of complete networks: An introduction to exponential random graph (p*) models for social networks. Social Networks, 29(2): 173-191 [pdf]