Importance of nonlinear regression models in the interpretation of data from COVID-19 in Colombia

Abstract

Introduction: Due to the harmful economic and social effects caused by the confinement of people, the Colombian government entities have planned an intelligent quarantine based on the interpretation of the behavior in the curve data, from which they affirm that it has shown a reduction during the last days.Objective: To highlight the importance of the analysis of non-linear correlation models and all the statistical inference procedures for the design of a mathematical model that allows the prediction of data based on the age of positive cases of COVID-19 in Colombia. Material and Methods: The daily results are based on the information obtained from the official website of the Colombian National Institute of Health. The total data are analyzed through the RKward free software (R Library). The aim of the analysis is to show the value of the correlation matrix, hypothesis test, r2 and the ideal correlation model, with which prediction is made.Results: With an R-squaredvalueof 0.9969 very close to 1 and a hypothesis test that guarantees the veracity of the alternative hypothesis, the ideal mathematical model that aligns the growth data of COVID-19 is quadratic.Conclusions: The quadratic model is positive and increasing as long as the number of infections continue to grow; therefore, it is not an ideal moment to speak of a flattening of the curve. If the growth is constant, the model could have an exponential trend.

Keywords

• predicción, curva de crecimiento, modelos de regresión.