Sequential Bayesian analysis of time-varying transmission in stratified epidemic renewal models

Date: Wednesday, Apr 15 2026
Time: 13:40 pm to 14:10 pm

Join the Webinar

Speaker

Dhorasso Junior Temfack Nguefack, PhD student, Trinity College Dublin, Ireland.

I am a PhD student in Statistics at Trinity College Dublin (TCD), working under the supervision of Dr. Jason Wyse in the School of Computer Science & Statistics.

My research focuses on sequential Bayesian inference for infectious disease modelling. I develop and apply particle filtering (Sequential Monte Carlo) methods for real-time estimation of transmission dynamics from epidemic data, with an emphasis on computationally efficient algorithms for state-space epidemic models.

My recent work includes developing sequential Monte Carlo squared methods for online inference in stochastic epidemic models and ensemble data assimilation approaches for sequential Bayesian inference.

Abstract

Age-specific differences in susceptibility, transmissibility, and social behavior drive heterogeneous epidemic dynamics that homogeneous models cannot capture. To address this, we develop a Bayesian age-stratified transmission model based on a semi-mechanistic renewal process, which links current infections to past incidence while accounting for population structure. The framework simultaneously reconstructs latent infection trajectories across age groups and infers time-varying transmissibility, enabling estimation of the instantaneous reproduction number. Inference is performed using a block sequential Monte Carlo (SMC) approach tailored for high-dimensional latent states, mitigating particle degeneracy by leveraging the weak coupling between age groups. We illustrate the method using COVID-19 data from Ireland, demonstrating the ability to infer age-specific transmissibility.

Category: Events