基于支持向量回归机的中国碳排放预测模型

选取人口、城镇化率、人均GDP、服务业增加值比重、单位GDP能耗、煤炭消费比例等6项影响因素作为自变量,运用支持向量回归机方法构建中国碳排放预测模型。以1980—2009年碳排放及影响因素数据为样本,通过训练、测试得到具有良好学习与推广能力的支持向量回归机模型。结合"十二五"规划,设置不同情境下影响因素预测值,对2010—2015年中国碳排放进行预测。预测结果表明,中国可适当降低GDP增速,不断优化能源结构,以确保碳减排目标的有效实现。

China’s carbon emissions prediction model based on support vector regression

Six influnce factors including population,urbanization rate,per capita GDP,added value proportion of service industry,per GDP energy consumption and coal consumption ratio were seleted as independent variables,and a model based on support vector regression(SVR) was established for predicting carbon emissions of China.Using the data of carbon emissions and influence factors from the year 1980 to 2009 as samples,the SVR model with good learning and generalization ability was established through training and testing.According to the 12th five-year program,prediction values of influence facors under different situations were set,and the carbon emissions of China from the year 2010 to 2015 were predicted.The results show that China can appropriately reduce GDP growth speed and constantly optimize energy structure so as to achieve carbon reduction target efficiently.