摘要
在对中国碳排放交易市场碳交易价格形成机制讨论的基础上,提出了预测指标体系。利用2017年1月1日—2018年9月30日广州碳交易市场碳交易价格数据和指标体系中各预测变量的数据,应用Lasso回归方法对变量进行筛选,建立灰色BP神经网络对碳交易价格进行预测。预测模型对于10期以内短期预测平均相对绝对误差(MAPE)小于4%,预测精度较高。
Based on the discussion of the price formation mechanism of China carbon emission trading market,the paper tried to propose the prediction index system of carbon trading price.Guangzhou market was chosen as the sample and data was collected from January 1,2017 to Sept-ember 30,2018.With Via Lasso regression it screened out the important factors and predicted the carbon trading price through the gray BP neural network model.It shows that the model's accuracy is relatively high and the average relative absolute error(MAPE)is less than 4%for the short-term prediction in the period of no more than 10.
作者
金林
马忠芸
王红红
Jin Lin;Ma Zhongyun;Wang Honghong(School of Statistics and Mathematics,Zhongnan University of Economics and Law,Wuhan 430073,China)
出处
《河北环境工程学院学报》
CAS
2020年第1期27-32,41,共7页
Journal of Hebei University of Environmental Engineering
