中文摘要：

本文采用中国1978-2010年的万元GDP能耗数据,分别构建了中国万元GDP能耗的普通自回归预测模型与分位自回归预测模型,对未来中国能源强度的变动趋势进行预测。研究发现,分位自回归模型的预测精度比普通自回归模型更高。因此,以分位自回归模型来对中国未来几年的能源强度进行预测。相比2015年,2020年中国万元GDP能耗将下降10.61%,低于政府提出的15%的目标,中国在节能道路上仍需努力。

英文摘要：

This paper adopts the data of the year 1978-2010 about energy consumption per unit GDP to build two forecast models with ordinary least square method and quantile auto regression method respectively. We forecast China’s energy intensity in the future and fund that the forecast results coming from quantile auto regression model are more precise. So we used the quantile auto regression model to forecast the energy intensity of China in the latter years and concluded that the energy intensity in 2020 would reduce by 10.61% comparing with the year 2015. It is lower than the target that the Chinese Government put forward. Thus, there will still be a long way for China to save energy.