Utilizando Shapley Additive Explanations para analisar características epidemiológicas da Covid-19.

Abstract

The COVID-19 pandemic resulted in millions of deaths globally, making the understanding of epidemiological characteristics a priority. In this context, institutions like Fiocruz provided essential data for analysis, and the use of Machine Learning and Explainable Artificial Intelligence techniques emerged as a strategy to interpret patterns in these data. This article, through a quantitative exploratory approach, presents a comparative analysis of the epidemiological characteristics of COVID-19, from 2020 to 2021. For this analysis, an eXplainable Artificial Intelligence technique, SHAP, was applied to Machine Learning models. Finally, it was possible to conduct the proposed analysis and identify both similarities and dissimilarities between the studied periods.