Reference Prior Development for Exponential Random Graph Model

• Activity Number: 503
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Sayan Chakraborty*+ and Tapabrata Maiti
• Companies: Michigan State University and Michigan State University
• Keywords: ERGM ; Reference Priors ; Jeffreys's Priors

Abstract

Due to the current advancement in social media, an individual has opportunity to propagate information to the largest number of peers ever in the history. Recently, the scientific interest has led to the rapid development of statistical models for social network structure. Exponential Random Graph Model (ERGM) is very popular in this context. Standard Bayesian techniques may not work well for this model. In this paper, we develop reference prior distributions for the model parameters and discussed the techniques for dealing the normalizing constant.