基于GIOWA算子的我国碳排放量的组合预测研究

Combined Forecast Research on Carbon Emission in China Based on GIOWA Operator

基于1989—2018年中国碳排放量数据,采用多元线性回归、Hlot-Winters非季节指数平滑、ARIMA模型3种单项预测模型对我国碳排放量进行预测,鉴于单项预测模型的局限性,基于误差平方和最小的最优性准则,建立广义诱导有序加权平均(GIOWA)的组合预测模型,并对模型的有效性进行评价。结果表明:组合预测模型优于单项预测模型,验证了组合预测模型的有效性;未来5年,我国碳排放量处于上升趋势,而碳排放强度呈下降趋势。

Based on China's carbon emissions data from 1989 to 2018,with multiple linear regression model,Hlot-Winters non-seasonal exponential smoothing,and ARIMA model,we predict China's carbon emissions.In view of the limitation of the single-term prediction model,the paper establishes a generalized induced ordered weighted average(GIOWA)combination prediction model based on the optimality criterion of the minimum squared error and the minimum criterion.And then,we evaluates its effectiveness.The results show that:(1)the combined forecasting model is superior to the single forecasting model,which verifies the effectiveness of the combined forecasting model;(2)In the next five years,China's carbon emissions are on an upward trend,but the intensity of carbon emissions is on a downward trend.