David Hunter

I work on statistical models for social networks.

Statistical models for networks, with applications to disease dynamics

(with Mark Handcock, Martina Morris, and Steve Goodreau)

To understand how diseases spread through "typical" social networks (e.g. how STDs and HIV spread through sexual networks), it is important to understand what "typical" means. Statistically speaking, we aim to use observed network data to develop plausible probability models for network formation. We are developing the mathematical tools to do this.

statnet: An R package for statistical network modeling

(with Mark Handcock, Carter Butts, Steve Goodreau, and Martina Morris)

We aim to find probability models that describe network data well. We can then use these models to explore the dynamics of disease spread. We have written statnet — a publicly available package for the R statistical software environment — to:

• identify the best-fitting model from a candidate class of models using maximum likelihood estimation
• simulate test networks that retain the essential features of the observed data.

We can then investigate:

• how disease epidemics spread across these simulated networks
• how the network is affected by the disease.