Adjustment of time series models with intervention to predict gasoline consumption in Brazil

Authors

Keywords:

trend, seasonality, SARIMA model;, intervention, forecasts

Abstract

The gasoline consumption in Brazil has undergone considerable fluctuations over
the years due to several factors, including the launch of flex-fuel vehicles that can be moved by
gasoline, ethanol or a mixture of the two fuels. This work had as main objective to analyze
the behavior of the average gasoline consumption in thousand barrels/day in Brazil through time
series analysis, verify the effect of seasonality, trend and intervention and making forecasts from
the fitted models of Box Jenkins. The data regarding the average monthly consumption of ga-
soline in barrels/day, collected during the period january 1979 to april 2012 were obtained from
the database of the Institute of Applied Economic Research (IPEA). Were identified in the study
series on the presence of trend and seasonal components, and models from the Box Jenkins,
SARIMA models that were adjusted appropriately to the data. According to the criterion for se-
lection of models, AIC (Akaike information Criterion), BIC (Bayesian Information Criterion)
and MSE (Mean Square Error of Prediction), SARIMA model with the intervention was more appropriate for making forecasts.

Author Biographies

Nádia Giaretta Biase, Universidade Federal de Uberlândia

Faculdade de Matemática-FAMAT Área: Estatística

Maria Imaculdada de Sousa Silva, Universidade Federal de Uberlândia

Faculdade de Matemática-FAMAT Área: Estatística

References

BOX, G. P.; JENKINS, G. M. Time series analysis, forecasting and control. San Francisco: Holden-Day, 1976.

FERRAZ, M. I. F.; SÁFDI, T.; LAGE, G. Uso de modelos de séries temporais na previsão de séries de precipitação pluviais mensais no Município de Lavras-MG. Revista Brasileira de Agrometeorologia, v. 7, n.2, p. 259-267, 1999.

GONÇALVES, L. R. Modelagem de séries representativas do setor energético brasileiro, 2007. 106p. Dissertação (Mestrado em Estatística e Experimentação Agropecuária)-Universidade Federal de Lavras, Lavras, 2007.

INSTITUTO DE PESQUISA ECONÔMICA APLICADA.

Disponível em: . Acesso em: 20 jun. 2012.

MORETTIN, P. A; TOLOI, C. M. C. Análise de séries temporais. São Paulo: Edgard Blücher, 2006. 538p.

MURADÁSNETO, M. V.; CARVALHO, P. L. C.; SÁFADI, T. Análise preditiva e comparativa para séries de combustíveis no Município de Lavras. Revista Científica Symposium, v.6, n.1, p.55-61, 2008.

PRIESTLEY, M.B. Spectral analysis and time series. volume1: Univariate series. volume2: Multivariate series, Prediction and control. New York : Academic Press. 1989. 890 p.

SCHÜNEMANN, L. A demanda de gasolina automotiva no Brasil: o impacto nas elasticidades de curto e longo prazo da expansão do GNV e dos carros flex, 2007. 91p. Dissertação (Mestrado Profissionalizante em Economia) – Faculdade de Economia e Finanças IBMEC, Rio de Janeiro, 2007.

Published

31-12-2013

How to Cite

Biase, N. G., & de Sousa Silva, M. I. (2013). Adjustment of time series models with intervention to predict gasoline consumption in Brazil. Sigmae, 2(1), 23–33. Retrieved from https://publicacoes.unifal-mg.edu.br/revistas/index.php/sigmae/article/view/47

Issue

Section

Applied Statistics