Inferring key epidemiological parameters and transmission dynamics of COVID-19 based on a modified SEIR model

¹ Department of Mathematics, Shanghai Normal University, Shanghai 200234, P.R. China.
² School of Information Engineering, Zhengzhou University, Zhengzhou 450052, P.R. China.
³ International Research Center for Neurointelligence (WPI-IRCN), The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan.

^* Corresponding author: qguo@shnu.edu.cn

Received: 6 May 2020
Accepted: 23 November 2020

Abstract

This study aims to establish a model-based framework for inferring key transmission characteristics of the newly emerging outbreak of the coronavirus disease 2019 (COVID-19), especially the epidemic dynamics under quarantine conditions. Inspired by the shifting therapeutic levels and capacity at different stages of the COVID-19 pandemic, we propose a modified SEIR model with a two-phase removal rate of quarantined hosts undergoing continuously tunable transition. We employ the Markov Chain Monte Carlo (MCMC) approach for inferring and forecasting the epidemiological dynamics from the publicly available surveillance reports. The effectiveness of a short-term prediction is illustrated by adopting the data sets from 10 demographic regions including Chinese mainland and South Korea. In the surveillance period, the average R₀ ranges from 1.74 to 3.28, and the median of the mean latent period does not exceed 10 days across the surveillance regions.