Prof Lam Ho — “Efficient Bayesian Methods for Detecting Trajectory Shifts in Stochastic Models of Infectious Disease”
The trajectory of an infectious disease epidemic may change over time due to policy interventions, the emergence of new strains, or accumulating population immunity. In this talk, I will present our ongoing project on developing Efficient Bayesian Methods for Detecting Trajectory Shifts in Stochastic Models of Infectious Disease. A primary challenge of stochastic models is that likelihood calculations for prevalence count data have traditionally been considered computationally intractable, rendering these models difficult to scale for large outbreaks like the COVID-19 pandemic. To overcome this limitation, our method introduces a Poisson approximation scheme that significantly reduces computational cost. Crucially, we demonstrate that this approximation scheme preserves information, maintaining accuracy comparable to scenarios where the data are continuously observed.
Prof Denise Lievesley — “Protecting the integrity of official statistics”
Denise will draw on her experience as Director of Statistics in UNESCO, working in many countries of the world, and also on her more recent appointment to review the UK Statistical system, in order to discuss threats to the integrity of official statistics. She will talk about the various codes and principles which exist to protect the independence of the National Statistics Institutes and will consider how those of us outside of the public service can act as critical friends.
Book now