4TH International Congress on Technology - Engineering & SCIENCE - Kuala Lumpur - Malaysia (2017-08-05)

Forecasting And Decomposing Energy Consumption Using The Gm (1, 1) Model And The Information Entropy: The Case Of Hunan Province Of China

An important economic index which reflects the industrial development of a city or a country is energy consumption. It is very significant for us to predict future energy consumption accurately. Forecasting energy consumption by conventional methods usually requires a large sample size. However, the date collected on energy consumption are often a small sample size or non-normal. Hence, to obtain detailed understanding of the future amount of Hunan’s energy consumption in the coming years, the GM(1,1) model, three fitting models, and NAR time series neural network method were used to forecast the total energy consumption in Hunan. Using information entropy model and neural network model to conduct the domain decomposition of consumption under the control of total energy consumption. Moreover, five methods were used to predict the amount of the total energy consumption during the period of present to 2020. Then use two models to decompose the total energy consumption. Contrasting analysis different models, it is found that the data obtained by grey prediction and information entropy is more consistent with the actual situation. Finally, it is found that how to improve energy efficiency and find alternative energy sources, will become a major problem facing the future of Hunan.
Keywords: energy consumption, energy decomposition, GM (1, 1) model, neural network model, information entropy model

Chuanchang Li, Guiyu Xiao, Jian Chen