Predictive model to identify indexes by country that affect the mortality rate from COVID-19

Pereáñez J.A.G.; Arango D.A.G.; Camarillo C.H.; Fernández F.H. · RISTI - Revista Iberica de Sistemas e Tecnologias de Informacao (2021)

Resumen

The incursion of the coronavirus contagion phenomenon and its subsequent classification as a pandemic has been unforeseen and significantly impacts the indicators of all countries worldwide. The economic reports that measure the consequences of the measures of the different governments to reduce the rate of infections in the context of the pandemic, show a significant reduction in income and employment that ultimately results in a decrease in the quality of life of the population. In order to understand a little more the consequences that this situation of high uncertainty brings with it, this article presents the results of the construction of a predictive analysis based on neural networks by means of which 27653 data from the daily report of dates between the January 22 and July 10, 2020 and which were extracted from the coronanet project in which, through the interaction of indices by country, the order of importance of predictors of the virus mortality rate was identified. Among the results obtained, in addition to obtaining a model with a precision of 77.5%, it is highlighted that although the type of measure adopted by governments is an incident factor in the increase in mortality per country, they are specific characteristics such as the transparency index and the globalization index which allow a predictive analysis of mortality.

Instituciones: Universidad Nacional Toribio Rodríguez de Mendoza de Amazonas