Modelo sazonal fracionário auto-regressivo integrado média móvel com volatilidade: estudo de previsão das concentrações de material particulado inalável

Abstract

Studies on time series have been widely used to analyze environmental data. This study evaluates the daily concentration of inhalable particulate matter (PM10) for the Region of Grande Vitória, ES, Brazil, for the years 2003 and 2004, using SARFIMA-GARCHprocesses, models that capture the properties of long-term dependence, seasonality andvolatility Conditional variable over time. To verify the effectiveness of the methodology was carried out studies for the comparison of prediction and forecasting models to theusual literature, and the results showed that the proposed model, SARFIMA-GARCHbest captured variability of the data in terms of average square error.