Scholarlink

<p>To mitigate the greenhouse effect and promote green, sustainable development, this study evaluated the carbon emission efficiency (CEE) of 27 cities in the Yangtze River Delta urban agglomeration (YRDUA), China. An interpretable machine learning model, based on the SHapley Additive exPlanations (SHAP) method and optimized with ant colony optimization (ACO), was employed to identify the effects of relevant factors on CEE. Furthermore, the SHAP-based results were innovatively applied to Moran’s I test to examine the spatial clustering of these effects. The findings indicate that (1) the spatial clustering pattern of CEE within the YRDUA became increasingly evident during the study period, accompanied by pronounced interprovincial disparities and increasingly differentiated provincial and municipal distribution patterns. (2) Human activities partially offset the positive effects of some factors on CEE, while technological progress and energy efficiency improvements significantly enhanced the positive effects of others. (3) The impact of population density on CEE exhibited significant spatial clustering effects, forming three distinct clusters due to the interaction between urban development on different stages and other influencing factors. By developing an integrated analytical framework that combines efficiency evaluation, interpretable machine learning, and spatial analysis, this study advances CEE research and provides targeted insights for urban decarbonization.</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

Exploring the spatial correlation of factor impacts on urban carbon emission efficiency using DEA-XGBoost-SHAP methods: evidence from the Yangtze River Delta, China

  • 期刊论文

  • 发表日期：

    2025年11月3日

查看全文
  • Junfei Chen,
  • Yichi Jia,
  • Yaning Yang,
  • Yuqi Zhang

摘要

To mitigate the greenhouse effect and promote green, sustainable development, this study evaluated the carbon emission efficiency (CEE) of 27 cities in the Yangtze River Delta urban agglomeration (YRDUA), China. An interpretable machine learning model, based on the SHapley Additive exPlanations (SHAP) method and optimized with ant colony optimization (ACO), was employed to identify the effects of relevant factors on CEE. Furthermore, the SHAP-based results were innovatively applied to Moran’s I test to examine the spatial clustering of these effects. The findings indicate that (1) the spatial clustering pattern of CEE within the YRDUA became increasingly evident during the study period, accompanied by pronounced interprovincial disparities and increasingly differentiated provincial and municipal distribution patterns. (2) Human activities partially offset the positive effects of some factors on CEE, while technological progress and energy efficiency improvements significantly enhanced the positive effects of others. (3) The impact of population density on CEE exhibited significant spatial clustering effects, forming three distinct clusters due to the interaction between urban development on different stages and other influencing factors. By developing an integrated analytical framework that combines efficiency evaluation, interpretable machine learning, and spatial analysis, this study advances CEE research and provides targeted insights for urban decarbonization.