• Media type: E-Article
  • Title: Forecasting China's energy intensity by using an improved DVCGM (1, N) model considering the hysteresis effect
  • Contributor: Meng, Zhaosu; Liu, Xiaotong; Yin, Kedong; Li, Xuemei; Guo, Xinchang
  • Published: Emerald, 2021
  • Published in: Grey Systems: Theory and Application, 11 (2021) 3, Seite 372-393
  • Language: English
  • DOI: 10.1108/gs-02-2020-0022
  • ISSN: 2043-9377
  • Origination:
  • Footnote:
  • Description: PurposeThe purpose of this paper is to examine the effectiveness of an improved dummy variables control grey model (DVCGM) considering the hysteresis effect of government policies in China's energy intensity (EI) forecasting.Design/methodology/approachEnergy consumption is considered as an important driver of economic development. China has introduced policies those aim at the optimization of energy structure and EI. In this study, EI is forecasted by an improved DVCGM, considering the hysteresis effect of energy-saving policies of the government. A nonlinear optimization method based on particle swarm optimization (PSO) algorithm is constructed to calculate the hysteresis parameter. A one-step rolling mechanism is applied to provide input data of the prediction model. Grey model (GM) (1, N), DVCGM (1, N) and ARIMA model are applied to test the accuracy of the improved DVCGM (1, N) model prediction.FindingsThe results show that the improved DVCGM provides reliable results and works well in simulation and predictions using multivariable data in small sample size and time-lag virtual variable. Accordingly, the improved DVCGM notes the hysteresis effect of government policies and significantly improves the prediction accuracy of China's EI than the other three models.Originality/valueThis study estimates the EI considering the hysteresis effect of energy-saving policies in China by using an improved DVCGM. The main contribution of this paper is to propose a model to estimate EI, considering the hysteresis effect of energy-saving policies and improve forecasting accuracy.