基于ARIMA和BP神经网络组合模型的我国碳排放强度预测

FORECAST OF INTENSITY OF CARBON EMISSION TO CHINA BASED ON BP NEURAL NETWORK AND ARIMA COMBINED MODEL

预测我国碳排放强度的长期变动趋势,对国家进行宏观经济管理和节能减排工作具有重要的参考价值。运用深入分析自回归移动平均模型和神经网络的特性,并在此基础上建立ARIMA模型和BP神经网络组合模型,将碳排放强度的时间序列的数据结构分解为线性和非线性残差部分,对我国碳排放强度的变化趋势进行了综合分析与预测。结果显示:今后10a我国碳排放强度总体是逐步下降的,但到2020年我国碳排放强度仅比2005年下降34%,比我国政府提出碳排放强度下降40%～45%的目标还有一定的差距。因此,要在2020年实现我国碳排放强度目标,必须要调整宏观经济政策,采取各种政策措施以实现目标。

Abstract： Forecasting long-term intensity of carbon emissions in China has important significance for policy makers to macro-economy management and Energy-Saving Emission Reduction Efficiency. Based on analysis of the autoregressive integrated moving average （ARIMA） and neural networks （NN） models,this paper presented an ensemble approach to the intensity of carbon emission time series forecasting which inte- grated ARIMA with NN. The time series was considered of a linear autocorrelation structure and nonlinear structure, and then the change trend of intensity of carbon emissions was analyzed and predicted. The fore- cast results indicate that the carbon intensity is gradually declined in the next ten years, but the intensity of carbon emissions in 2020 decreases only by 34% based on the 2005 level. The Chinese government pro- claimed a mitigation target which proposed that intensity of carbon emissions in 2020 would be reduced by 40%--45% based on the 2005 level. Therefore,we have to adjust macroeconomic policy and take all kinds of policy measures to achieve the goal.