Compartmentalized mathematical model to predict future number of active cases and deaths of COVID-19

Abstract

In December 2019, China reported a series of atypical pneumonia cases caused by a new Coronavirus, called COVID-19. In response to the rapid global dissemination of the virus, on the 11th of Mars, the World Health Organization (WHO) has declared the outbreak a pandemic. In light of this situation, this paper intends to analyze and improve the current SEIR models to better represent the behavior of the COVID-19 and accurately predict the outcome of the pandemic in a given social, economic and political scenario. We present a novel generalized Susceptible-Exposed-Infected-Recovered (SEIR) compartmental model and test it using a global optimization algorithm with data collected from the WHO. Our main results were: (a) our model was able to accurately fit the data of all countries tested (b) it is possible to predict values for one week ahead with errors in the order of 15% for the number of cases and 30% in the number of deaths for all countries; (c) predictions are better for countries where the active cases curve already reached the maximum; the error being in the order of 10% in the number of cases and 20% in the number of deaths; (d) for countries where the active curve is still growing, different optimization solutions can be found that fit the data; so, to predict future behavior in this scenarios some of the model coefficients should be estimated from outside sources or based on generalized results from other countries according to their health policies of social distance, quarantining and case test and tracing.

Estimation of parameters for a humidity-dependent compartmental model of the COVID-19 outbreak

“…Genetic Algorithms (GA) have been successfully utilized to optimize the parameters of epidemiological models for the Severe Acute Respiratory Syndrome (SARS) ( Yan & Zou, 2008 ), cholera ( Akman & Schaefer, 2015 ; Akman, Corby & Schaefer, 2016 ) and recently for the SARS-CoV-2 outbreak ( Neto et al, 2020 ; Yousefpour, Jahanshahi & Bekiros, 2020 ; Kozio, Stanisawski & Bialic, 2020 ).…”

“…An inappropriate model will perform poorly even with the best possible parameter settings, same way as the the most appropriate model with badly chosen parameter values. Newer studies, such as Zhou et al (2020) , Liu et al (2020a) , Neto et al (2020) , and Kozio, Stanisawski & Bialic (2020) , validated the proposed modells on data, in order to verify the parameter estimations and prediction of the trend of SARS-CoV-2 infections.…”

Modelling a novel Coronavirus (COVID-19): A stochastic SEIR-HCD approach, with real-time parameter estimation & forecasting for Scotland

“…Examples of further deterministic and statistical COVID-19 SEIR models incorporating hospitalisation, intensive care and deaths can be found in [13,14,15]. However, again, all of these models use parameter estimates that are constant for all time, based upon figures within the literature.…”

Applying Benford’s law to COVID-19 data: The case of the European Union

“…Due to the rapid contagiousness of the virus (Hafeez et al, 2020), nearly every country employed measures against the virus’ spread, such as national lockdowns and restrictions of outdoor activities. The pandemic showed that statistical and machine learning modelling can potentially predict the number of new cases or deaths for a given country (Cássaro & Pires, 2020; Niazkar & Niazkar, 2020; Neto et al, 2020). The accurate forecast of the infection curve can facilitate government’s measures towards the suppression of the growth rate.…”