Adjustment of non-linear models to Covid-19 notifications in the municipality of Palmas – Tocantins

Abstract

Understanding how the transmissibility of a disease occurs is important to establish the necessary measures to combat it. Thus, the objective is to identify which non-linear model, between the models of Gompertz, Von Bertallanfy and Logistic, would better adjust the curve of prevalence of cases of Covid-19 of the city of Palmas-TO. As well as analyzing data from case reports and deaths, in variables such as gender and age group. Data will be used until September 2021, found on the SESAU-TO website, the notifications are confirmed through the RT-PCR test. The formulas of the models will be compiled and represented in graphs. To define the best model will be developed an AI, whose lowest value indicates the best fit. We can observe that although the highest rate of affection was in adults, the highest lethality was in the elderly. In relation to the variable gender, the highest rate of affection was in the female sex, and of lethality in the male sex. We can conclude that the model that presented the best fit to the notifications was the one from Gompertz, with a lower OI value. In addition, the estimated value found at the end of the period observed in the Gompertz model was the most similar to the number of notifications in Palmas among the models analyzed. It was also observed that Von Bertalanffy’s model presented a low AI, being also a curve suitable for Palmas. However, as for the Logistic method curve, it was the most adjusted at the beginning of the period, but as the mutations progressed, it did not become effective for the municipality.