资源开发与市场2025，Vol.41Issue(9)：1281-1295,15.DOI:10.3969/j.issn.1005-8141.2025.09.001
基于多源数据与深度学习的中国碳市场价格行为研究
Research on price behavior of China's carbon market based on multi-source data and deep learning
摘要
Abstract
This paper studied the behavior of carbon prices by compiling a low-carbon attention index and constructing a hybrid model based on multi-source data,using historical carbon market price data from Hubei Province as an example.The results indicated that:① The short-term excessive volatility and long-range correlation of carbon prices,as well as the right-fat-tail distribution of the series,were the reasons for the multifractality in the carbon market.② There existed a bidirectional causal relationship between car-bon prices and factors such as the economy,energy,carbon management,carbon resources,and low-carbon attention.③ Short-to me-dium-term changes in carbon prices were most influenced by past levels,while medium-term economic factors were the primary con-tributors to carbon price changes.In the long-term equilibrium state,low-carbon attention had the greatest impact on carbon prices,followed by carbon resources,then carbon management(with a negative impact),and finally energy and the economy.关键词
碳市场/多源数据/MF DFA/informer模型/混合模型/低碳关注指数Key words
carbon market/multi-source data/MF DFA/Informer model/hybrid model/low-carbon attention index分类
资源环境引用本文复制引用
张陶新,欧思丽,黄蕊..基于多源数据与深度学习的中国碳市场价格行为研究[J].资源开发与市场,2025,41(9):1281-1295,15.基金项目
教育部人文社科一般项目(编号:23YJA910001) （编号:23YJA910001）
广州华商学院科研团队项目(编号:2021HSKT03) （编号:2021HSKT03）
广东省重点建设学科科研能力提升项目(编号:2022ZDJS127). （编号:2022ZDJS127）
