Coronavirus Epidemic Spreading Prediction combining random walk with susceptible–exposed–infected–recovered–deceased (SEIRD) model

Abstract

Since the breakout of COVID-19 pandemic in December 2019, effective modelling of the spreading of the virus has become an essential reference for the epidemic controlling. In a bid to solve the problem of Epidemic prediction, susceptible-exposed-infected-recovered (SEIR) model are widely applied. However, this model seems lack the ability to handle random events which may occur during the spreading of the pandemic and the ability to simulate the pandemic spreading between different subdivided regions. Therefore, we propose an early version of susceptible–exposed–infected–recovered–deceased (SEIRD) model that combines the classic compartmental concepts of SEIRD and the random walk methodology to forecast COVID-19 in real time. Specifically, this study will focus on improvement of the exposed–infected part of SEIRD model. First, the exposed–infected section of SEIRD model will be applied to each subdivided regions separated. Then, instead of entering infected–recovered part directly, the infected of each district will be selected and sent to linked districts by random walk system to mimic the commuting and irregular trips between regions. Eventually, after the re-distribution of infected patients, the model will enter the infected–recovered section. This argued model adopt the SEIRD model to forecasting of virus spreading between small regions and taking irregular moving of citizens into consideration via random walk system, thus provide an effective reference for countries which aim to respond to the post-epidemic era.