文章编号： 1672-5328（2026）01-0104-09

付铎1，李瑞敏2，戚扬1，曹瑾鑫1, 3
(1. 内蒙古大学交通学院，内蒙古自治区呼和浩特010000；2. 清华大学土木工程系，北京100084；3. 内蒙古科 学技术研究院，内蒙古自治区呼和浩特010000)

摘要： 城市轨道交通网络客运强度的影响机理较为复杂，从城市层面分析社会经济与城市轨道交通 系统属性对其作用机制，对于促进城市轨道交通系统的可持续发展具有重要意义。基于2023 年中 国大陆地区44 个开通城市轨道交通城市的实际数据，从城市轨道交通系统属性、城市轨道交通可 达性、社会经济特征以及城市交通系统特征4 个方面选取了10 个变量，综合运用主成分分析、极限 梯度提升模型和SHAP解释器，系统探究了城市轨道交通网络客运强度的关键影响因素及其作用机 理。结果表明：各影响因素对网络客运强度的作用存在显著阈值效应。其中，运营规模、城市交通 拥堵指数和换乘车站比例是主要正向影响因素；平均站间距与网络客运强度呈负相关；公共汽车线 网密度、道路网密度等因素的影响机制较为复杂，需结合具体城市情况进行分析。

关键词： 城市轨道交通；网络客运强度；影响因素；XGBoost模型；SHAP分析；阈值效应

中图分类号： U491.1+2

文献标识码：A

Factors Influencing Passenger Volume Intensity in Urban Rail Transit Networks

Fu Duo1, Li Ruimin2, Qi Yang1, Cao Jinxin1, 3
(1. School of Transport, Inner Mongolia University, Hohhot Inner Mongolia Autonomous Region 010000, China; 2. School of Civil Engineering, Tsinghua University, Beijing 100084, China; 3. Inner Mongolia Academy of Science and Technology, Hohhot Inner Mongolia Autonomous Region 010000, China)

Abstract： The mechanisms influencing passenger volume intensity in urban rail transit networks are highly complex. Analyzing the effects of socioeconomic conditions and urban rail transit system attributes at the city level is significant for promoting the sustainable development of urban rail transit systems. Based on real data from 44 cities in Chinese mainland with operational urban rail transit systems in 2023, this study selects ten variables across four dimensions, including urban rail transit system attributes, rail transit accessibility, socioeconomic characteristics, and urban transportation system features. Principal component analysis, an Extreme Gradient Boosting (XGBoost) model, and the SHAP explainer are used to systematically investigate the key factors affecting passenger volume intensity in urban rail transit networks and their underlying mechanisms. The results indicate that the influencing factors exhibit pronounced threshold effects on network passenger volume intensity. Among them, operational scale, the urban traffic congestion index, and the proportion of transfer stations emerge as the primary positive influencing factors. Average station spacing shows a negative correlation with network passenger volume intensity. The influence mechanisms of factors, such as bus transit network density and road network density, are more complex and require analysis tailored to the specific conditions of individual cities.

Keywords： urban rail transit; network passenger volume intensity; influencing factors; XGBoost model; SHAP analysis; threshold effects