Scholarlink

Background <p>As global aging accelerates, community public spaces (CPS) are increasingly recognized as vital for promoting healthy aging. However, existing research often employs linear analytical methods or focuses on single health dimensions, overlooking how CPS features are non-linearly and interactively related to multiple aspects of elderly well-being.</p> Methods <p>This study analyzed data from 2,508 older adults in Shanghai to examine the associations between 13 CPS environmental features and physical, mental, and social health. An explainable machine learning approach, incorporating CatBoost and SHAP, was used to identify key correlates and patterns of interaction among features.</p> Results <p>Destination accessibility, public transport, residential density, and walking infrastructure were positively associated with all health dimensions, while traffic complexity and crime rate showed consistent negative associations. SHAP interaction analysis revealed that health benefits often emerge from synergistic combinations of features, rather than from the effects of individual features in isolation.</p> Conclusions <p>This study provides new insights into the multidimensional and non-linear relationships between CPS features and elderly health. The findings offer empirical evidence to guide community planning and intervention strategies that foster health-supportive environments for aging populations.</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 complex associations between community public spaces and healthy aging: an explainable analysis using catboost and SHAP

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年7月2日

查看全文
  • Man Zhang,
  • Tao Shen,
  • Ying Li,
  • Qianqian Li,
  • Yongqi Lou

摘要

Background

As global aging accelerates, community public spaces (CPS) are increasingly recognized as vital for promoting healthy aging. However, existing research often employs linear analytical methods or focuses on single health dimensions, overlooking how CPS features are non-linearly and interactively related to multiple aspects of elderly well-being.

Methods

This study analyzed data from 2,508 older adults in Shanghai to examine the associations between 13 CPS environmental features and physical, mental, and social health. An explainable machine learning approach, incorporating CatBoost and SHAP, was used to identify key correlates and patterns of interaction among features.

Results

Destination accessibility, public transport, residential density, and walking infrastructure were positively associated with all health dimensions, while traffic complexity and crime rate showed consistent negative associations. SHAP interaction analysis revealed that health benefits often emerge from synergistic combinations of features, rather than from the effects of individual features in isolation.

Conclusions

This study provides new insights into the multidimensional and non-linear relationships between CPS features and elderly health. The findings offer empirical evidence to guide community planning and intervention strategies that foster health-supportive environments for aging populations.