中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊
重庆交通大学学报（自然科学版）

重庆交通大学学报（自然科学版） ›› 2025, Vol. 44 ›› Issue (8): 59-65.DOI: 10.3969/j.issn.1674-0696.2025.08.08

• 交通+大数据人工智能 • 上一篇    

考虑交叉分类特性的职住地建成环境对通勤方式选择的影响

尹超英1，曹昱泽1，金玮2，邵春福3   

  1. (1. 南京林业大学 汽车与交通工程学院, 江苏 南京 210037；2. 江苏高速公路联网营运管理有限公司, 江苏 南京 210049；3. 新疆大学 交通运输工程学院, 新疆 乌鲁木齐 830017)
  • 收稿日期:2024-08-19 修回日期:2025-01-14 发布日期:2025-09-05
  • 作者简介:尹超英(1989—)，女，山西五寨人，副教授，博士，主要从事出行行为及居民幸福感建模方面的研究。E-mail: cyyin@njfu.edu.cn 通信作者：邵春福(1957—)，男，河北沧州人，教授，博士，主要从事交通规划、通管理和智能交通方面的研究。E-mail: cfshao@xju.edu.cn
  • 基金资助:
    国家自然科学基金项目（72204114）；教育部人文社科项目（22YJC630191）；中国博士后科学基金项目（2023M731705）

Influence of Built Environment of Residential Area and Workplace on Commuting Mode Choice Considering Cross-Classification Characteristics

YIN Chaoying1，CAO Yuze1，JIN Wei2，SHAO Chunfu3   

  1. (1. College of Automobile and Traffic Engineering, Nanjing Forestry University, Nanjing 210037, Jiangsu, China; 2. Jiangsu Expressway Network Operation Management Co., Ltd., Nanjing 210049, Jiangsu, China; 3. School of Traffic and Transportation Engineering, Xinjiang University, Urumqi 830017, Xinjiang, China)
  • Received:2024-08-19 Revised:2025-01-14 Published:2025-09-05

PDF

摘要： 为克服传统多层模型难以同时对居住地和工作地空间异质性建模问题，利用长春市居民出行调查数据，构建贝叶斯交叉分类离散选择模型，从空间视角探究建成环境对居民通勤方式选择的影响。结果表明：较不考虑居住地建成环境的多层二项Logistics模型和不考虑空间异质性的单层二项Logistics模型而言，考虑空间异质性的贝叶斯交叉分类离散选择模型拟合效果更优，由居住地和工作地空间异质性导致的居民通勤方式选择变异程度分别占总变异的22.5%和11.6%；在控制个体社会经济属性的基础上，居住地和工作地的建成环境特性大多与小汽车通勤使用呈现显著相关性，土地利用混合度对居民出行方式选择影响呈显著负相关，职住地到CBD的距离对小汽车通勤影响呈显著正相关；交叉口密度方面，工作地交叉口密度影响不显著，而居住地交叉口密度影响呈显著负相关。研究结论可为城市建成环境的合理配置与出行需求的正确引导提供理论基础。

关键词: 交通运输工程；建成环境；通勤方式；空间异质性；交叉分类模型

Abstract: In order to overcome the problem that the traditional multi-level model was difficult to model the spatial heterogeneity of residential area and workplace at the same time, a Bayesian cross-classification discrete choice model was constructed by use of survey data on residents travel in Changchun city, to explore the impact of the built environment on residents choice of commuting mode from a spatial perspective. The results show that compared to multi-level binomial Logistics models without considering the built environment of residential areas and binomial Logistics models without considering spatial heterogeneity, the Bayesian cross-classification discrete choice model with considering spatial heterogeneity has better fitting performance. The degree of variation in residents commuting mode choices caused by spatial heterogeneities in their residential area and workplace accounts for 22.5% and 11.6% of the total variation, respectively. On the basis of controlling individual socio-economic attributes, the built environment characteristics of residential area and workplace are mostly significantly correlated with the use of cars for commuting, and the land use mixing degree has a significant negative correlation with residents travel mode choices. The distance from the residential area and workplace to the CBD has a significant positive correlation with the impact of commuting by car. In terms of intersection density, the impact of workplace intersection density is not significant, while the impact of residential area intersection density is significantly negatively correlated. The research results can provide theoretical basis for the reasonable allocation of urban built environment and correct guidance of travel demand.

Key words: traffic and transportation engineering; built environment; commuting mode; spatial heterogeneity; cross-classification model

中图分类号: