生态学报2025，Vol.45Issue(11)：5229-5245,17.DOI:10.20103/j.stxb.202411242877
基于人工神经网络的数字经济碳减排及其条件效应
Research on carbon emission reduction and its conditional effects of digital economy based on Artificial Neural Networks
摘要
Abstract
How the digital economy can promote dual control of carbon emissions and intensity within the framework of sustainable development is attracting widespread attention and in-depth exploration from all sectors of society.With the increasing integration of digital technologies into various industries,understanding their potential to contribute to environmental sustainability is crucial.To gain a deeper understanding of the intrinsic connection between the digital economy and carbon emissions,this paper innovatively applies an artificial neural network model.This model systematically explores the net effects and conditional dependencies of the digital economy on total carbon emissions,carbon intensity,and per capita carbon emissions,providing a comprehensive analysis of how digital transformation impacts carbon reduction efforts.The results show that:①The digital economy has played a crucial role in reducing total carbon emissions,with the reduction rate accelerating as economic development levels increase.Its impact on carbon intensity and per capita carbon emissions follows distinct patterns,exhibiting a negatively skewed inverted"U"shape and an"S"-shaped trend,respectively.Notably,the significant reduction effects on both carbon intensity and per capita emissions become evident only when the digital economy reaches a sufficiently high threshold,highlighting the importance of sustained digital advancement.②Compared to traditional econometric methods,the artificial neural network model demonstrates significantly superior accuracy in capturing the complex relationships between digital economy development and carbon emissions.When fitting the net effects on total carbon emissions,carbon intensity,and per capita emissions,the model's accuracy increased by 65.58％,67.86％,and 56.57％,respectively,demonstrating its enhanced predictive power.③The impact of the digital economy on emissions varies significantly across different urban contexts,including city size,administrative level,and the presence of digital economy pilot zones.The most substantial emission reductions occur in super-large and mega-cities,although further efforts are needed to optimize reductions in carbon intensity and per capita emissions.In cities with higher administrative levels,digital economy development helps lower carbon intensity and per capita emissions,but does not significantly affect total emissions.Pilot zones show strong institutional advantages,particularly in green development and emission reduction efforts.④The net effect of the digital economy on carbon emissions is influenced by various conditions,including economic development,foreign direct investment,innovation,and technology focus.These factors play a crucial role in shaping how digital economy development impacts carbon emissions,suggesting that the dual control of emissions can be effectively promoted through strategic advancements in these four aspects.关键词
数字经济/碳排放总量/碳排放强度/人均碳排放量/条件效应/人工神经网络Key words
digital economy/total carbon emissions/carbon emission intensity/per capita carbon emissions/conditioned effect/artificial neural network引用本文复制引用
吴伟平,刘雨宁,苏乐言,张俊狮,刘帅跇..基于人工神经网络的数字经济碳减排及其条件效应[J].生态学报,2025,45(11):5229-5245,17.基金项目
国家社会科学基金后期资助项目(24FJYB051) （24FJYB051）
国家社会科学基金重大项目(23&ZD067) （23&ZD067）
湖南省自然科学基金面上项目(2024JJ5117) （2024JJ5117）
湖南省教育厅科学研究重点项目(23A0487) （23A0487）
