Credibility Intervals for the Reproduction Number of the Covid-19 Pandemic Using Proximal Langevin Samplers - ENS de Lyon - École normale supérieure de Lyon Accéder directement au contenu
Proceedings/Recueil Des Communications 2023 31th European Signal Processing Conference (EUSIPCO) Année : 2023

Credibility Intervals for the Reproduction Number of the Covid-19 Pandemic Using Proximal Langevin Samplers

Résumé

Monitoring the time evolution of the intensity of the Covid-19 pandemic within the pandemic and despite the limited quality of the data is both crucial and challenging. In the context of a pandemic, providing a level of confidence in the estimation of epidemiological indicators is essential to inform decision makers. The present work proposes a Bayesian estimation of the pandemic Covid-19 reproduction number and of the denoised reported counts through credibility intervals. The challenge stems from an epidemiological Bayesian model robust to errors in reported counts, which yields a non differentiable a posteriori log-density. Four different Hastings-Metropolis algorithms combining Langevin approaches and proximal operators are compared on a toy example; the most efficient one is plugged into a Metropolis-within-Gibbs algorithm performing a credibility intervals-based estimation of Covid-19 pandemic indicators, exemplified for several countries worldwide.
Fichier principal
Vignette du fichier
23ICASSP_CovidICv2.pdf (1017.73 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03902144 , version 1 (15-12-2022)
hal-03902144 , version 2 (31-07-2023)

Identifiants

  • HAL Id : hal-03902144 , version 1

Citer

Patrice Abry, Gersende Fort, Barbara Pascal, Nelly Pustelnik. Credibility Intervals for the Reproduction Number of the Covid-19 Pandemic Using Proximal Langevin Samplers. 2023 31th European Signal Processing Conference (EUSIPCO), 2023. ⟨hal-03902144v1⟩
209 Consultations
77 Téléchargements

Partager

Gmail Facebook X LinkedIn More