COVID-19 lethality in Brazilian States using information theory quantifiers

Abstract

In this paper, we presented an overview diagnosis consider the time series of daily deaths by COVID-19 in the Brazilian States using Bandt & Pompe method (BPM) to estimate the Information Theory quantifiers, more specifically the Permutation entropy (H s ) and the Fisher information measure (F s ). Based on the Information Theory quantifiers, we build up the Shannon-Fisher causality plane (SFCP) to promote insights into the COVID-19 temporal evolution inherent in the phenomenology associated with the number of daily deaths well as their respective locations along the SFCP. Moreover, we apply H s and F s to elaborate on the rank of the Brazilian States’ real situation, considering the number of daily death due to COVID-19 based on the complexity hierarchy. The Brazilian States that are located in the middle region of the two-dimensional plane (H s x F s ), such as Amapá (AP), Roraima (RO), Acre (AC), and Tocantins (TO) are characterized by a less entropic and low disorder, which implies in high predictability of the COVID-19 lethality. While, the Brazilian States that are located in the lower-right region, such as Ceará (CE), Bahia (BA), Pernambuco (PE), and Rio de Janeiro (RJ), are characterized by high entropy and high disorder, which leads to low predictability of the COVID-19 lethality. Given this, our results provide empirical evidence that the permutation entropy is a powerful approach to predicting infectious diseases. Dynamic monitoring of permutation entropy can help policymakers to take more or less restrictive measures to combat COVID-19.