China’s energy consumption prediction considering error correction based on decompose–ensemble method

Abstract

Energy consumption forecasting plays a vital role in rational energy and economic planning formulation for a country or an area around the world. A novel forecasting method based on error correction and decompose-ensemble strategy combined with linear regression model (LR) and triple exponential smoothing model (TESM) is proposed in this study. Firstly, original energy consumption is decomposed into the main trend subseries and non-stationary errors subseries by use of LR model. And the mainly trend subseries is forecasted by LR. Then, the non-stationary subseries (forecasting errors of LR) is decomposed into several intrinsic mode functions (high-frequency subseries, middle-frequency subseries and low-frequency subseries) and residual subseries by empirical mode decomposition (EMD). With respect to their different dynamic changing features and influenced factors, each intrinsic mode functions subseries and residual subseries are forecasted, respectively by TESM. Finally, the prediction of energy consumption is obtained by summing the trend subseries prediction results and these errors subseries prediction results. Forecasting results prove that error correction is a useful strategy for improve the forecasting performance. Most of all, the origin complex errors correction forecasting problem has been resolved into some simple forecasting problem based on decompose-ensemble strategy, which have better forecasting performance, compared with individual models (LR, auto regression model (AR) and TESM), traditional error correction method, combination models (which is developed by using of LR, AR and TESM based on equal weight method, entropy weight method and optimal weight method). The proposed novel method can provide accurate and reliable forecasting results, which is a feasible forecasting approach for China’s annual energy consumption. By use of the proposed forecasting method considering error correction based on EMD decompose-ensemble strategy combined with LR and TESM, China’s energy consumption in 2021 will increase to 484,555.30 ten thousand standard tons coal equivalent (tce), and the average annual growth of the coming 5 years is 2.135%. Since China’s energy consumption is still on its growing process, China should pay more attention to change its economic development model from energy-intensive economy to low-carbon economy.