Predicción de infectados por Covid-19 en el Perú por el modelo de media móvil integrada autorregresiva

• Autor: Aro Huanacuni, Alex Youn
• Formato: artículo
• Fecha de Publicación: 2022
• Institución: Universidad Nacional Jorge Basadre Grohmann
• Repositorio: Revistas - Universidad Nacional Jorge Basadre Grohmann
• Lenguaje: español
• OAI Identifier: oai:revistas.unjbg.edu.pe:article/1237
• Enlace del recurso: https://revistas.unjbg.edu.pe/index.php/cyd/article/view/1237
• Nivel de acceso: acceso abierto
• Materia: Infectados; Muertes; Modelo ARIMA; Predicción

During the outbreak of the Covid-19 virus, several researchers have studied various mathematical models for predicting infections and deaths, as well as the rate of virus transmission. At present, the virus is still active with some variants and it is very important to know its behavior in order to develop effective actions to control the current and future situation. In the research, we obtained predictions of cumulative Covid-19 infections for 38 days from December 23, 2021, using data recorded in the World Health Organization (WHO) for Peru and training the autoregressive integrated moving average (ARIMA) model in the software Python 3. The most optimal models obtained with real data test and according to EMPA and R2 are ARIMA(3,0,1) in the prediction of infected with EMPA=0.178 and R2=0.804 and ARIMA(3,1,1), with EMPA= 0.243 and R2=0.579, for prediction of deaths.