Pandemic and covid-19: a mathematical analysis of the (un) predictability of cases in the state of amazon

The pandemic of COVID-19 began in May 2020, as declared worldwide by the World Health Organization (WHO), with approximately 589,680,368 cases and 6,436,519 deaths from the disease being confirmed by the year 2022. In Amazon, up to August 14, 2022, 14,249 Covid-19 deaths have been recorded, 342 this year alone. This is an epidemiological study of a quantitative nature, in the application of mathematical models for event prediction and management in health services. The process of mathematical modeling and data mining was composed of the following steps: library, data preparation, test setup, validation, and results, in the epidemiological weeks comprising the period from 2020 to 2021. The period was composed of 90 epidemiological weeks, total of 22,247 new cases, in the epidemiological week equivalent to 3,178 new cases per day. The 90-week period was maintained, with a total of 14,120 deaths; after applying the moving average, we worked with 1,052 deaths registered at week 56. The mathematical models GBM and KKNN, used in this study, showed a significant difference between the predicted and the actual observed number of cases and deaths from Covid-19 in the second wave of the pandemic in Amazon.