Scholarlink

<p>Under the pressing backdrop of global climate change, residents’ carbon emissions (RCE) are a key part of urban carbon emissions and a critical area for achieving low-carbon transitions. This study innovatively integrates random forest algorithms with the Geographically and Temporally Weighted Regression (GTWR) model to analyze the spatiotemporal heterogeneity of RCE across 287 Chinese cities from 2013 to 2022. It also develops an ESDA-SDE model via map visualization to explore the spatiotemporal evolution of urban RCE, and applies a random forest model to screen for influential factors, optimizing a GTWR model to analyze the spatiotemporal heterogeneity of these factors. The findings indicate that while China’s urban RCE continues to rise in total volume, its growth rate is gradually decreasing. Spatially, RCE exhibits a “high in the east, low in the west” pattern, with a standard deviation ellipse showing the RCE centroid shifting southwestward, reflecting faster RCE growth in the western and southern regions and suggesting significant potential for future carbon emissions growth. The influencing factors display marked spatiotemporal heterogeneity. Over time, economic development, industrial structure, urbanization, and population density positively affect RCE, while technological investment and policy intensity contribute to carbon reduction with growing influence. Spatially, the industrial structure significantly impacts the eastern region, economic development and technological investment are more influential in the northeastern and Inner Mongolia regions, carbon reduction policies are more effective in the northeastern and Shandong regions, and technological investment is more substantial in the northern and eastern regions.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimizing of random forest algorithm to analyze the spatiotemporal heterogeneity of urban resident carbon emissions in China

  • 期刊论文

  • 发表日期：

    2025年11月26日

查看全文
  • Jixiang Zhang,
  • Yu Zhao,
  • Sui Li,
  • Xiaoyang Duan,
  • Panpan Wang,
  • Tianping Bi

摘要

Under the pressing backdrop of global climate change, residents’ carbon emissions (RCE) are a key part of urban carbon emissions and a critical area for achieving low-carbon transitions. This study innovatively integrates random forest algorithms with the Geographically and Temporally Weighted Regression (GTWR) model to analyze the spatiotemporal heterogeneity of RCE across 287 Chinese cities from 2013 to 2022. It also develops an ESDA-SDE model via map visualization to explore the spatiotemporal evolution of urban RCE, and applies a random forest model to screen for influential factors, optimizing a GTWR model to analyze the spatiotemporal heterogeneity of these factors. The findings indicate that while China’s urban RCE continues to rise in total volume, its growth rate is gradually decreasing. Spatially, RCE exhibits a “high in the east, low in the west” pattern, with a standard deviation ellipse showing the RCE centroid shifting southwestward, reflecting faster RCE growth in the western and southern regions and suggesting significant potential for future carbon emissions growth. The influencing factors display marked spatiotemporal heterogeneity. Over time, economic development, industrial structure, urbanization, and population density positively affect RCE, while technological investment and policy intensity contribute to carbon reduction with growing influence. Spatially, the industrial structure significantly impacts the eastern region, economic development and technological investment are more influential in the northeastern and Inner Mongolia regions, carbon reduction policies are more effective in the northeastern and Shandong regions, and technological investment is more substantial in the northern and eastern regions.