Machine Learning (ML) algorithms: Prediction of COVID-19 cases in Bolivia

The economic system is going through uncertainty, resilience due to the pandemic, in the face of this adversity it is decisive to estimate the behaviors of the Covid-19 case scenarios, for strategic decision-making in public policies. The aim of this article is to examine the implications of machine learning (ML) algorithms on the accuracy of the prediction of Covid-19 cases. The methodology is based on predictive statistics and analysis of ML algorithms. The result of the evidence determines at a comparative level that the ARIMA and RANDOM FOREST algorithms do not have good precision or data fit, so their future forecasts are not fulfilled, on the other hand the PROPHET algorithm behaves regularly, as an intermediate indicator in the comparison of prediction with numbers of cases in real time, within that perspective are the predictive algorithms that have the best test precision and are adjusted to the reality of the daily data of Covid-19 cases for Bolivia; GLMNET AND PROPHET W / XGBOOT ERRORS in the two simulated phases prior to the presentation of this article, therefore it could be concluded that the GLMNET and PROPHET W / XGBOOT ERRORS predictive models are better algorithms to predict Covid-19 case scenarios.

• Autor principal: Ruiz Otondo, Juan (author)
• Formato: article
• Lenguaje: spa
• Publicado: 2021
• Materias: Machine Learning; Covid-19; Predicción
• Acceso en línea: https://estudioseconomicos.bce.fin.ec/index.php/RevistaCE/article/view/339