Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Performance evaluation of forecasting models based on time series and machine learning techniques: an application to light fuel consumption in Brazil

Performance evaluation of forecasting models based on time series and machine learning... Fuel demand forecast is a fundamental tool to guide private planning actions and public policies aim to guarantee energy supply. This paper aims to evaluate different forecasting methods to project the consumption of light fuels in Brazil (fuel used by vehicles with internal combustion engine).Design/methodology/approachEight different methods were implemented, besides of ensemble learning technics that combine the different models. The evaluation was carried out based on the forecast error for a forecast horizon of 3, 6 and 12 months.FindingsThe statistical tests performed indicated the superiority of the evaluated models compared to a naive forecasting method. As the forecast horizon increase, the heterogeneity between the accuracy of the models becomes evident and the classification by performance becomes easier. Furthermore, for 12 months forecast, it was found methods that outperform, with statistical significance, the SARIMA method, that is widely used. Even with an unprecedented event, such as the COVID-19 crisis, the results proved to be robust.Practical implicationsSome regulation instruments in Brazilian fuel market requires the forecast of light fuel consumption to better deal with supply and environment issues. In that context, the level of accuracy reached allows the use of these models as tools to assist public and private agents that operate in this market.Originality/valueThe study seeks to fill a gap in the literature on the Brazilian light fuel market. In addition, the methodological strategy adopted assesses projection models from different areas of knowledge using a robust evaluation procedure. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Energy Sector Management Emerald Publishing

Performance evaluation of forecasting models based on time series and machine learning techniques: an application to light fuel consumption in Brazil

Download PDF
23 pages

Loading next page...
 
/lp/emerald-publishing/performance-evaluation-of-forecasting-models-based-on-time-series-and-Aj9d6kQdeO
Publisher
Emerald Publishing
Copyright
© Emerald Publishing Limited
ISSN
1750-6220
eISSN
1750-6220
DOI
10.1108/ijesm-02-2021-0009
Publisher site
See Article on Publisher Site

Abstract

Fuel demand forecast is a fundamental tool to guide private planning actions and public policies aim to guarantee energy supply. This paper aims to evaluate different forecasting methods to project the consumption of light fuels in Brazil (fuel used by vehicles with internal combustion engine).Design/methodology/approachEight different methods were implemented, besides of ensemble learning technics that combine the different models. The evaluation was carried out based on the forecast error for a forecast horizon of 3, 6 and 12 months.FindingsThe statistical tests performed indicated the superiority of the evaluated models compared to a naive forecasting method. As the forecast horizon increase, the heterogeneity between the accuracy of the models becomes evident and the classification by performance becomes easier. Furthermore, for 12 months forecast, it was found methods that outperform, with statistical significance, the SARIMA method, that is widely used. Even with an unprecedented event, such as the COVID-19 crisis, the results proved to be robust.Practical implicationsSome regulation instruments in Brazilian fuel market requires the forecast of light fuel consumption to better deal with supply and environment issues. In that context, the level of accuracy reached allows the use of these models as tools to assist public and private agents that operate in this market.Originality/valueThe study seeks to fill a gap in the literature on the Brazilian light fuel market. In addition, the methodological strategy adopted assesses projection models from different areas of knowledge using a robust evaluation procedure.

Journal

International Journal of Energy Sector Management – Emerald Publishing

Published: May 11, 2022

Keywords: Co-integration; Forecasting; Time series analysis; Biofuels; Econometric; Demand forecasting; Autoregressive; Neural networks; Fuzzy-logic model; Demand-side management; Gasoline; Liquid fuels; Fuel demand; Forecasting methods; Time series; Machine learning; Forecast evaluation

References

  • Consumption statistics
  • A review on applications of ANN and SVM for building electrical energy consumption forecasting
    Ahmad, A.S.; Hassan, M.Y.; Abdullah, M.P.; Rahman, H.A.; Hussin, F.; Abdullah, H.; Saidur, R.
  • Short-run, long-run and cross elasticities of gasoline demand in Brazil
    Alves, D.C.O.; Bueno, R.L.S.
  • Resolution no. 67, of 9.12.2011
  • Resolution no. 819, of 6.5.2020
  • Fuel sales in 2020 dropped 5.97% compared to 2019 due to the pandemic
  • Statistical data
  • Time series management system
  • Ethanol consumption in Brazil: empirical facts based on persistence, seasonality and breaks
    Barros, C.P.; Gil-Alana, L.A.; Wanke, P.
  • A note on the validity of cross-validation for evaluating autoregressive time series prediction
    Bergmeir, C.; Hyndman, R.J.; Koo, B.
  • Random forests
    Breiman, L.
  • A demanda por gasolina no Brasil: uma análise utilizando técnicas de co-integração
    Burnquist, H.L.; Bacchi, M.R.P.
  • Biofuels policies and fuel demand elasticities in Brazil
    Cardoso, L.C.B.; Bittencourt, M.V.L.; Litt, W.H.; Irwin, E.G.
  • Modelo de previsão de demanda por combustíveis automotivos no Brasil
    Castro, G.R.
  • Monthly electricity demand forecasting based on a weighted evolving fuzzy neural network approach
    Chang, P.C.; Fan, C.Y.; Lin, J.J.
  • Short-term load forecasting using a kernel-based support vector regression combination model
    Che, J.; Wang, J.
  • Short-term electrical load forecasting using the support vector regression (SVR) model to calculate the demand response baseline for office buildings
    Chen, Y.; Xu, P.; Chu, Y.; Li, W.; Wu, Y.; Ni, L.; Bao, Y.; Wang, K.
  • Combining forecasts: a review and annotated bibliography
    Clemen, R.T.
  • Resolution no. 17, of 8.6.2017
  • Resolution no. 8, of 18.8.2020
  • Testing normality in the presence of outliers
    Coin, D.
  • Modeling fuel demand in Brazil: the importance of the flex-fuel vehicle
    Costa, C.C.; Burnquist, H.L.; Souza, M.J.P.; Valdez, C.
  • Public consultation no. 94 of June 5, 2020
  • A review on time series forecasting techniques for building energy consumption
    Deb, C.; Zhang, F.; Yang, J.; Lee, S.E.; Shah, K.W.
  • Forecasting methods in energy planning models
    Debnath, K.B.; Mourshed, M.
  • Statistical comparisons of classifiers over multiple data sets
    Demsar, J.
  • Determining the order of differencing in autoregressive processes
    Dickey, D.A.; Pantula, S.G.
  • The economics of Brazil’s ethanol-sugar markets, mandates, and tax exemptions
    Drabik, D.; Gorter, H.; Just, D.R.; Timilsina, G.R.
  • Flex-fuel vehicle adoption and dynamics of ethanol prices: lessons from Brazil
    Du, X.; Carriquiry, M.A.
  • Short-Term load forecasting using random forests. Intelligent systems’ 2014
    Dudek, G.; Filev, D.; Jabłkowski, J.; Kacprzyk, J.; Krawczak, M.; Popchev, I.; Rutkowski, L.; Sgurev, V.; Sotirova, E.; Szynkarczyk, P.; Zadrozny, S.
  • Mixture of mlp-experts for trend forecasting of time series: a case study of the Tehran stock exchange
    Ebrahimpour, R.; Nikoo, H.; Masoudnia, S.; Yousefi, M.R.; Ghaemi, M.S.
  • Testing autocorrelation in a system perspective testing autocorrelation
    Edgerton, D.; Shukur, G.
  • Efficient tests for an autoregressive unit root
    Elliott, G.; Rothenberg, T.J.; Stock, J.H.
  • Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation
    Engle, R.F.
  • Machine learning based switching model for electricity load forecasting
    Fan, S.; Chen, L.; Lee, W.J.
  • Forecasting fuel ethanol consumption in Brazil by time series models: 2006-2012
    Figueira, S.R.; Burnquist, H.L.; Bacchi, M.R.P.
  • Demand for automotive fuel in Brazil: an empirical analysis using cointegration techniques
    Figueira, S.R.F.; Santos, D.F.L.; Borges, A.C.G.; Sanches, A.; Campos, A.L.S.
  • Ethanol demand under the flex-fuel technology regime in Brazil
    Freitas, L.C.; Kaneko, S.
  • PM10 concentration forecasting in the metropolitan area of Oviedo (Northern Spain) using models based on SVM, MLP, VARMA and ARIMA: a case study
    García Nieto, P.J.; Lasheras, F.S.; García-Gonzalo, E.; Cos Juez, F.J.
  • An analysis of the impact of Flex-Fuel vehicles on fuel consumption in Brazil, applying cointegration and the Kalman filter
    Gomez, J.M.A.; Legey, L.F.L.
  • Electric load forecasting by support vector model
    Hong, W.-C.
  • Short-term forecasting of the wave energy flux: analogues, random forests, and physics-based models
    Ibarra-Berastegi, G.; Saénz, J.; Esnaola, G.; Ezcurra, A.; Ulazia, A.
  • Population statistics
  • Another driver of the Brazilian fuel ethanol supply chain: the consumers’ preferences
    Isabella, G.; Yu, A.S.O.; Silva, A.M.; Pegetti, A.L.
  • Statistical analysis of cointegration vectors
    Johansen, S.
  • Estimation and hypothesis testing of cointegration vectors in Gaussian vector
    Johansen, S.
  • Forecasting electricity demand in Thailand with an artificial neural network approach
    Kandananond, K.
  • Comparison of ARIMA and random forest time series models for prediction of avian influenza H5N1 outbreaks
    Kane, M.J.; Price, N.; Scotch, M.; Rabinowitz, P.
  • Modeling and prediction of turkey’s electricity consumption using artificial neural networks
    Kavaklioglu, K.; Ceylan, H.; Ozturk, H.K.; Canyurt, O.E.
  • An artificial neural network (p, d, q) model for timeseries forecasting
    Khashei, M.; Bijari, M.
  • HyFIS: adaptive neuro-fuzzy inference systems e their application to nonlinear dynamical systems
    Kim, J.; Kasabov, N.
  • Turkey’s short-term gross annual electricity demand forecast by fuzzy logic approach
    Kucukali, S.; Baris, K.
  • Testing the null hypothesis of stationarity against the alternative of a unit root: how sure are we that economic time series have a unit root?
    Kwiatkowski, D.; Phillips, P.C.B.; Schmidt, P.; Shin, Y.
  • Day-ahead load forecast using random forest and expert input selection
    Lahouar, A.; Ben Hadj Slama, J.
  • Seasonality and the order of integration for consumption
    Osbom, D.R.; Chui, A.P.L.; Smith, J.P.; Birchenhall, C.R.
  • Consumer choice between ethanol and gasoline: lessons from Brazil and Sweden
    Pacini, H.; Silveira, S.
  • Long term distribution demand forecasting using neuro fuzzy computations
    Padmakumari, K.; Mohandas, K.P.; Thiruvengadam, S.
  • Hybrid ellipsoidal fuzzy systems in forecasting regional electricity loads
    Pai, P.F.
  • Forecasting electricity demand on short, medium and long time scales using neural networks
    Ringwood, J.V.; Bofelli, D.; Murray, F.T.
  • Light fuel demand and public policies in Brazil, 2003-2013
    Rodrigues, L.; Bacchi, M.R.P.
  • Analyzing light fuel demand elasticities in Brazil using cointegration techniques
    Rodrigues, L.; Bacchi, M.R.P.
  • Economic-financial performance of the Brazilian sugarcane energy industry: an empirical evaluation using financial ratio, cluster and discriminant analysis
    Rodrigues, L.; Rodrigues, L.
  • Build it, but will they come? Evidence from consumer choice between gasoline and sugarcane ethanol
    Salvo, A.; Huse, C.
  • Um modelo integrado econométrico+insumo-produto Para previsão de longo prazo da demanda de combustíveis no Brasil
    Santiago, F.S.; Mattos, R.S.; Perobelli, F.S.
  • Fuel demand in Brazil in a dynamic panel data approach
    Santos, G.F.
  • Flexible-fuel automobiles and CO2 emissions in Brazil: parametric and semiparametric analysis using panel data
    Santos, A.S.; Gilio, L.; Halmenschlager, V.; Diniza, T.B.; Almeida, A.N.
  • An analysis of variance test for normality (complete samples)
    Shapiro, S.S.; Wilk, M.B.
  • Comparison of different post HOC tests
    Shingala, M.C.
  • Turkey’s net energy consumption
    Sözen, A.; Arcaklioğlu, E.; Özkaymak, M.
  • Energy models for demand forecasting – a review
    Suganthi, L.; Samuel, A.A.
  • Brazilian automobile and light fuel fleet (Otto cycle)
  • Generating fuzzy rules by learning from examples
    Wang, L.X.; Mendel, J.M.
  • Um modelo híbrido para previsão de curto prazo da demanda de gasolina automotiva no Brasil
    Zanini, A.; Souza, R.C.; Pedreira, C.E.