Statistics seminar

​Dootika Vats, Indian Institute of Technology, Kanpur: Revisiting the Gelman-Rubin Diagnostic

​Abstract: Gelman and Rubin's (1992) convergence diagnostic is one of the most popular methods for terminating a Markov chain Monte Carlo (MCMC) sampler. Since the seminal paper, researchers have developed sophisticated methods of variance estimation for Monte Carlo averages. We show that this class of estimators find immediate use in the Gelman-Rubin statistic, a connection not established in the literature before. We incorporate these estimators to upgrade both the univariate and multivariate Gelman-Rubin statistics, leading to increased stability in MCMC termination time. An immediate advantage is that our new Gelman-Rubin statistic can also be calculated for a single chain. In addition, we establish a relationship between the Gelman-Rubin statistic and effective sample size. Leveraging this relationship, we develop a principled cutoff criterion for the Gelman-Rubin statistic. Finally, we demonstrate the utility of our improved diagnostic via an example. This work is joint with Christina Knudson, University of St. Thomas, Minnesota.

About the speaker: Dootika Vats is Assistan Professor at the Indian Instute of Technology in Kanpur. She is the 2021 recipient of the ISBA "Blackwell-Rosenbluth Award". She works in the general area of Bayesian computation and specifically focus on Markov chain Monte Carlo algorithms.
The seminar will take place in the room MVL15 and also on zoom (a password is needed: email
​Organiser: Umberto Picchini ( Please contact me if you wish to be informed of future seminars.
Kategori Seminarium
Plats: MVL15 and online
Tid: 2021-11-09 14:15
Sluttid: 2021-11-09 15:15

Sidansvarig Publicerad: fr 05 nov 2021.