Análise de séries temporais aplicada ao programa de coleta seletiva no município de Uberlândia (MG)

Abstract

The present study makes use of time series analysis techniques in a data set referring to the selective collection in the city of Uberlândia (MG) during the years 2011 to 2019. The time series explored is a monthly amount of recyclable material collected in the city by the selective collection service, available on the Uberlândia City Hall website. The Box e Jenkins and exponential smoothing techniques were used to build models that capture the information contained in series, in order to evaluate the behavior of the variable analyzed over time, as well as predict future values. Forecasting can contribute to the policymaking process responsible for better waste management. When applying the Box e Jenkins model, a SARIMA forecasting model was adopted, which considers both the trend and seasonality of the series, with the one with the lowest Akaike Information Criterion (AIC) being selected. For the purpose of comparing two forecasting techniques, the root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) were adopted. The Exponential Smoothing model had fewer errors, or better performance, than SARIMA model.