Análise da eficiência do modelo sir quando confrontado com dados reais

Abstract

Considering the great importance of studying mathematical models that help understand the dynamics of transmission of infectious diseases, especially the SIR model, This work aims to investigate the efficiency of the SIR model applied to COVID-19 data in the state of Amazonas. To this end, studies were carried out on the SIR model, also on COVID-19 and, in addition, an algorithm was created in Python language to carry out the necessary analyses, comparing with real data provided by the Health Surveillance Foundation (FVS). As a result, we saw that the simple SIR model is not efficient compared to the real data provided by official bodies, since depending on the applications, while the model estimates very high peaks, the real data shows much smaller peaks. In this way, we can understand the discrepancy as being caused mainly by underreporting, since the simple SIR considers that all cases are diagnosed. Therefore, in this work we present an extension of the SIR model to also consider underreporting.