Inference for Social Networks Based on Epidemic Data
Open Access
- Author:
- Groendyke, Christophe
- Graduate Program:
- Statistics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- December 03, 2010
- Committee Members:
- David Hunter, Committee Chair/Co-Chair
Murali Haran, Committee Member
Debashis Ghosh, Committee Member
Padma Raghavan, Committee Member - Keywords:
- networks
ERGM
MCMC
Erdos-Renyi
measles - Abstract:
- In this dissertation, we explore the problem of performing inference for the parameters of a random graph model, which is used to describe the structure of contact relationships within a population, and those of a stochastic epidemic model, using data from an epidemic that is assumed to have spread through the population via this contact network. We employ a Bayesian framework and MCMC integration to make estimates of the joint posterior distribution of the model parameters. We show that this type of inference is indeed viable for larger and more complex data sets than those that have previously been considered in the literature; we also demonstrate how additional types of data can be incorporated into this inferential procedure. We greatly expand the class of network models used for this type of inference and show that in many scientifically interesting cases, we can not only distinguish the effects of the network model from those of the epidemic model, but that it is also possible to determine the relative importance of the various factors influencing contact network structure. We demonstrate some of the important aspects of our approach by studying a measles outbreak in Hagelloch, Germany in 1861 consisting of 188 affected individuals. We show that our approach yields some interesting insights into the nature of this outbreak and network structure that were absent from previous analyses. We provide an R package to carry out these analyses, which is available publicly on the Comprehensive R Archive Network (CRAN).