Aplicación web para la clasificación de pacientes de COVID-19 mediante algoritmos de aprendizaje automático

Abstract

Abstract: Due to the health crisis, health institutions need technological advances to detect and fight the covid-19 pandemic. Machine learning techniques can help identify infected patients and detect the spread of the virus. Furthermore, these techniques can be used to analyze available information from patients and help in the decision-making process in order to provide better care and disease control. In the present study, the application of machine learning techniques is proposed to classify patients with covid-19 based on their clinical characteristics. Three machine learning algorithms are evaluated using a dataset with information from confirmed covid-19 patients from a hospital. The results show that the random forest algorithm has a better performance, achieving 96% in the quality measure of precision, 95% sensitivity, 85% specificity, 99% AUC and 96% F-Score with the set balanced data. The performance of the algorithm is visualized in a web application developed according to the best practices of the DevOps framework.