Abstract
Forecasting energy consumption is not easy because of the nonlinear nature of the time series for energy consumptions, which cannot be accurately predicted by traditional forecasting methods. Therefore, a novel hybrid forecasting framework based on the ensemble empirical mode decomposition (EEMD) approach and a combination of individual forecasting models is proposed. The hybrid models include the autoregressive integrated moving average (ARIMA), the support vector regression (SVR), and the genetic algorithm (GA). The integrated framework, the so-called EEMD-ARIMA-GA-SVR, will be used to predict the primary energy consumption of an economy. An empirical study case based on the Taiwanese consumption of energy will be used to verify the feasibility of the proposed forecast framework. According to the empirical study results, the proposed hybrid framework is feasible. Compared with prediction results derived from other forecasting mechanisms, the proposed framework demonstrates better precisions, but such a hybrid system can also be seen as a basis for energy management and policy definition.
Highlights
Research on energy supply and demand has become critical since the 1973 oil crisis
To yield more accurate evaluation results, we further provide results being derived by mean absolute error (MAE) and mean square error (MSE) as references
The comparison results show that the hybrid models with ensemble empirical mode decomposition (EEMD) decomposition can significantly reduce overall forecasting errors
Summary
The average annual worldwide energy consumption grew due to the rapid economic growth of major economies. Over 90% of the world’s energy consumption comes from coal, oil, natural gas, and nuclear sources [1]. The understanding and prediction of energy consumption in general, and of a specific economy in particular, are critical from an economic perspective. Such energy predictions can help the government sector define relevant energy policies for sustainable development. The time series of the Taiwanese primary energy consumption (Figure 2) was adopted to verify the effectiveness of the prediction models. Annual primary energy consumption from 1965 to 2014 was provided by BP [1]. The time series was separated into two subsets where 90% (46 samples) of the dataset were chosen as the training set while the remaining 10% (4 samples) were selected as the test set for verifying the prediction efficiency
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
