Avaliação de Modelos de Aprendizado de Máquina para Triagem de Covid-19 com Base em Sintomas Autodeclarados.

Abstract

The Coronavirus outbreak, which began in 2019, rapidly overwhelmed healthcare The COVID-19 outbreak, which began in 2019, quickly overwhelmed healthcare systems, with a high number of severe cases requiring hospitalization and intensive care. Faced with resource shortages and difficulties in performing efficient triage, it became essential to seek solutions that could assist in the rapid identification of infected patients. This study aims to develop and evaluate Machine Learning models capable of predicting COVID-19 test results based on symptomatic data. For this purpose, a dataset provided by the Foundation for the Support of Science and Technology of the State of Pernambuco was used, containing records of symptoms and the corresponding test results collected during the pandemic. The proposal seeks to compare the performance of different classification algorithms, contributing to research focused on the application of artificial intelligence techniques in the healthcare field. Among the evaluated models, the Naive Bayes classifier achieved the best performance, with an F1-Score of 0.6964 on the test set, demonstrating greater predictive capability within the proposed scenario.