A Computational Approach to Finding SEIR Model Parameters that best Explain Infected and Recovered Time Series for SARS-CoV 2

Abstract

The novel SARS-CoV 2 coronavirus has grown to become a global pandemic. Since then, several approaches have been adopted and developed to provide insights into epidemic origins, worldwide dispersal and epidemiological history. The Susceptible, Exposed, Infected and Recovered (SEIR) models are among the widely used approaches to study the further progression of the pandemic. However, finding such model parameters remains a difficult task, especially in small geographical areas where details of the initial compartments and the model parameters deviates from global distributions. The main result of our paper is a meta-heuristic approach to find SEIR model parameters that best explains the current observed infected time series. Our approach, allows studying different future scenarios considering not only the most likely future, but a set of possible SEIR parameters that explains current epidemic trends. We show that there are several possible parameters sets of such models able to explain current epidemic trends and by studding them is possible to obtain insights into the future possible outcomes. We show that there are several possible parameters sets of such models able to explain current epidemic trends and by studding them is possible to obtain insights into the future possible outcomes.