刘超 刘泽润 刁弥

LIU Chao, LIU Zerun, DIAO Mi

既有研究发现城市密度会影响常见慢性病的发病率，但其理论机制和实证研究尚不充分，城市大数据的涌现和人工智能的发展为探索这一复杂影响提供了新的思路和方法。笔者以美国大都市区内典型城市人口普查区为例，分析人口、建成环境和活动的多维度密度因素同常见慢性病之间的关系，通过可视化方法探明关系趋势，利用多种机器学习和可解释性方法厘清关键影响要素、非线性关系、关键阈值、要素之间的交互作用和空间上的分异，得出不同慢病之间的共性和特性。研究结果验证了城市密度对慢性病的重要作用，发现了密度与慢性病之间的复杂关系，以及在不同地区之间的规律差异。这对建设健康城市、引导城市可持续发展、提升宜居水平具有重要意义。

Existed research has revealed that urban density would influence the incidence rates of common chronic diseases, but lack theoretical mechanism and empirical studies. The development of artificial intelligence (AI) and the urban big data provide new opportunities to explore original ideas and methods of complex correlations. Aiming at identifying the relationships between urban density and chronic diseases, this paper takes metropolitan census tracts in U.S. as the study areas and analyzes three dimensions of population density, built environment density, and activity density. With the help of data visualization, this paper initially investigates the trends between variables. By building several machine learning (ML) models and using interpretable methods, this paper then uncovers the key influencing features, describes non-linear relationships, finds key thresholds, explains the interactions of factors, and discovers spatial distinction. Finally, this paper summarizes similarities and differences of common chronic diseases. The results verify the significant impacts of urban density on chronic diseases, interpret the complicated relationships and regional disparity. The conclusions are of great importance for constructing health cities, guiding sustainable development, and promoting livability.

刘超，同济大学建筑与规划学院城市规划系，副教授，硕士生导师；智慧社区分试验中心，负责人。liuchao1020@tongji.edu.cn

刘泽润，纽约大学坦登工程学院城市系统，博士研究生

刁弥（通信作者），同济大学建筑与规划学院城市规划系，教授

• 收起