SSA NSW Branch seminar: Time-Varying Parameter Distributional Regression with Dynamic Global-Local Shrinkage Process Priors

The New South Wales Branch of the Statistical Society of Australia invites you to a seminar titled "Time-Varying Parameter Distributional Regression with Dynamic Global-Local Shrinkage Process Priors" delivered by Mattias Villani, Stockholm University.

Date: Friday 6 November 2026

Time: 4:30 pm AEDT

Venue: Room 4082, School of Mathematics and Statistics, UNSW Sydney, Kensington NSW 2052

Format: In person

Abstract: Time-Varying Parameter Distributional Regression with Dynamic Global-Local Shrinkage Process Priors

We consider regression models for exponential family distributions in which all distributional parameters are linked to covariates with time-varying regression coefficients. The coefficients follow dynamic global-local shrinkage process priors, which allow them to stay constant in some periods while moving rapidly or jumping in others. We develop a Gibbs sampling algorithm with three alternative updating steps for the coefficient trajectories. The samplers are based on Gaussian approximations and therefore target a perturbed posterior. Several strategies are proposed to improve sampling speed, robustness and accuracy. The degree of posterior perturbation is assessed by comparing with exact posterior inference from particle Gibbs with ancestor sampling. The coefficient innovations are defined so that they adapt automatically to the model geometry, the link functions, and the scale and covariance of the covariates. The approach is illustrated with a time-varying seasonal Poisson regression for UK bike rental demand and a Beta regression for US layoff proportions, with time-varying effects on both the mean and the precision. The algorithms will be made available in an open-source Julia package.

Presenter Bio: Mattias Villani, Stockholm University

Mattias Villani is Professor of Statistics at Stockholm University. His PhD is from Stockholm University in 2000, on the topic of Bayesian reduced rank models. He continued there as a Postdoctoral Researcher and Senior Lecturer. During this time, he also worked at the Central Bank of Sweden, where he developed models and methods for macroeconomic forecasting and policy analysis. In 2011, Mattias became Professor of Statistics at Linköping University, where he continued to work on Bayesian statistical methodology and computing, with applications to machine learning, neuroimaging, robotics, and transportation. While there, he founded the Statistics and Machine Learning Division and contributed to the integration of the two fields of Statistics and Machine Learning. In 2018, Mattias returned to Stockholm University, where he holds the Professor Chair of Statistics. He has been an Associate Editor for Econometrics and Statistics, and he currently serves as Associate Editor for Bayesian Analysis and Annals of Statistical Data Science.