Dr. Ick Hoon Jin

University of Texas, MD Anderson Cancer Center

Title : "Fitting Exponential Random Graph Models Using Varying Truncation Stochastic Approximation MCMC Algorithm".

Date and Time: February 10, 2014 - 11:00am to 12:00pm
Location: Cupples I, Room 113 (on the landing at the west side of our building)

Abstract: The exponential random graph model (ERGM) plays a major role in social network analysis. However, parameter estimation for the ERGM is a difficult problem due to model degeneracy as well as the intractability of its normalizing constant. The existing algorithms, such as Monte Carlo maximum likelihood estimation (MCMLE) and stochastic approximation, often fail for this problem in the presence of model degeneracy. In this seminar, I introduce the varying truncation stochastic approximation Markov chain Monte Carlo (SAMCMC) algorithm to tackle this problem. The varying truncation mechanism enables the algorithm to choose an appropriate starting point and an appropriate gain factor sequence, and thus to produce a reasonable parameter estimate for the ERGM even in the presence of model degeneracy. The numerical results indicate that the varying truncation SAMCMC algorithm can significantly outperform the MCMLE and stochastic approximation algorithms: for degenerate ERGMs, MCMLE! and stochastic approximation often fail to produce any reasonable parameter estimates, while SAMCMC can do; for non-degenerate ERGMs, SAMCMC can work as well as or better than MCMLE and stochastic approximation.
Key Words: Exponential Random Graph Model; Intractable Normalizing Constant; Model Degeneracy; Stochastic Approximation MCMC; Trajectory Averaging.