Autori: Pokharel, Gyanendra, Hossain, Shakhawat, Poitras, Colton
Titolo: Supervised classification of spatial epidemics incorporating infection time uncertainty
Periodico: Statistical methods & applications : Journal of the Italian Statistical Society
Anno: 2024 - Volume: 33 - Fascicolo: 2 - Pagina iniziale: 703 - Pagina finale: 722

Mechanistic models are key to providing reliable information for developing infectious disease control strategies. In general, these models are fitted in Bayesian Markov chain Monte Carlo (MCMC) frameworks that incorporate heterogeneities within a population. However, these frameworks have the major drawback of being computationally expensive. This problem is even more severe when the epidemic history is incomplete, such as unknown infection times. Instead of using the time-consuming Bayesian MCMC methods, this paper explores the use of supervised classification methods to analyze the infectious disease data incorporating infection time uncertainty. The epidemic generating models are classified based on summary statistics of epidemics as inputs. The validity of these methods is investigated by using simulated epidemic data and Tomato Spotted Wilt Virus (TSWV) data, accounting for unknown infectious periods and infection times of individuals. We show that these methods are capable of capturing biological characteristics of disease transmission dynamics when there is infection time uncertainty in infectious disease data.




SICI: 1618-2510(2024)33:2<703:SCOSEI>2.0.ZU;2-N

Esportazione dati in Refworks (solo per utenti abilitati)

Biblioteche ACNP che possiedono il periodico