| 198 | 0 | 3 |
| 下载次数 | 被引频次 | 阅读次数 |
【背景】随着地铁网络大规模建设与成网运营格局的不断完善，地铁客流量迅速增长，乘客出行需求与模式日益复杂多变，给地铁的运营管理带来新的挑战。【目标】依托手机信令数据连续追踪用户出行轨迹的优势，根据基站布设位置和辐射范围确定地铁站点内产生的信令数据，结合出行活动时间等关键阈值识别单次地铁出行，进而挖掘地铁出行模式，为优化地铁服务提供支撑。【方法】基于地铁网络拓扑模型并结合Dijkstra算法，重构乘客出行路径，进而获得全过程逐日出行数据，并采用两步分类方法挖掘乘客出行行为异质性，根据出行频次将用户分为高频用户和低频用户，从出行时间、空间和路径使用特征等维度提出时间规律性、典型出行、路径混合熵等指标，再使用K-means++聚类算法对高频和低频用户进一步细分。【数据】上海市2019年5月共包含448万名地铁用户产生的4亿条手机信令数据。【结论】提取到383万位用户的3009万次出行，18%的高频用户贡献了67%的出行，而82%的低频用户仅贡献了33%的出行。其中高频用户可分为单一路径依赖型通勤群体、路径选择灵活型通勤群体、非通勤目的日常出行群体3类；低频用户可分为商务出行群体、休闲娱乐出行群体、单日游或过境出行群体3类。研究成果可为优化地铁资源配置、制定精准营销策略以及提升地铁运行管理效率提供依据。
Abstract：[Background]With the large-scale construction and increasingly integrated operation of metro networks, metro ridership has grown rapidly. Passenger travel demands and patterns have become increasingly complex and diverse, presenting new challenges for metro operations and management. [Objective] Leveraging the advantages of mobile signaling data in continuously tracking users' travel trajectories, metro travel episodes are identified on the basis of the layout and coverage of base stations within metro stations, combined with thresholds like travel activity time. Next, typical metro travel patterns are mined to support metro optimization. [Method] Based on the metro network topology model and Dijkstra algorithm, passenger travel paths are reconstructed to obtain detailed daily travel records. A two-step classification method is then employed to uncover the heterogeneity of passenger travel behavior. Specifically, the users are divided into high-and low-frequency groups on the basis of their travel frequencies. Subsequently, using indicators such as temporal regularity, spatial distribution, and route utilization, the K-means++ clustering algorithm is employed to further refine the segmentation of each group. [Data] The study uses mobile signaling data from Shanghai in May 2019, which include 400 million signaling records generated by 4.48 million metro users. [Conclusions] The analysis extracts 30.09 million metro trips from 3.83 million users. Highfrequency users(18% of the total) contributed 67% of all trips, whereas low-frequency users(82%)accounted for only 33%. High-frequency users can be classified into three groups: commuters relying on a single route, commuters with flexible route choices, and regular users traveling for noncommuting purposes. Low-frequency users can be classified into three categories: business, leisure and entertainment, and single-day or transit travelers. The findings can inform resource allocation, targeted marketing strategies, and help improve operational efficiency in metro systems.
[1]中国城市轨道交通协会.城市轨道交通2023年度统计和分析报告[R].北京：中国城市轨道交通协会，2024.
[2]上海市统计局. 2023年上海市国民经济和社会发展统计公报[R].上海：上海市统计局，2024.
[3]STEENBRUGGEN J, TRANOS E, NIJKAMP P. Data from mobile phone operators:a tool for smarter cities[J].Telecommunications Policy, 2015, 39(3/4):335-346.
[4]吴华意,胡秋实,李锐,等.城市人口时空分布估计研究进展[J].测绘学报, 2022, 51(9):1827-1847.WU Huayi, HU Qiushi, LI Rui, et al. Research progress on spatio-temporal distribution estimation of urban population[J]. Acta Geodaetica et Cartographica Sinica, 2022,51(9):1827-1847.
[5]WANG Z, HE S Y, LEUNG Y. Applying mobile phone data to travel behaviour research:a literature review[J].Travel Behaviour and Society, 2018, 11:141-155.
[6]JIANG S, FERREIRA J, GONZALEZ M C. Activitybased human mobility patterns inferred from mobile phone data:a case study of Singapore[J]. IEEE Transactions on Big Data, 2017, 3(2):208-219.
[7]刘振国,齐崇楷,王江锋,等.数据驱动的城市群综合运输通道识别算法与特征分析[J].交通运输系统工程与信息, 2025, 25(3):73-84.LIU Zhenguo, QI Chongkai, WANG Jiangfeng, et al. Data-driven identification algorithm and feature analysis of integrated transport corridors in urban agglomerations[J].Journal of Transportation Systems Engineering and Information Technology, 2025,25(3):73-84.
[8]郑成龙,宋辞,陈洁.基于手机信令数据的北京市长时间工作现象空间特征分析[J].地球信息科学学报,2025, 27(6):1317-1331.ZHENG Chenglong, SONG Ci, CHEN Jie. An analysis of spatial characteristics of long working hours phenomenon in Beijing based on mobile signaling data[J]. Geo-Information Science, 2025,27(6):1317-1331.
[9]ZHONG G, YIN T, ZHANG J, et al. Characteristics analysis for travel behavior of transportation hub passengers using mobile phone data[J]. Transportation, 2019, 46(5):1713-1736.
[10]HUANG H, CHENG Y, WEIBEL R. Transport mode detection based on mobile phone network data:a systematic review[J]. Transportation Research Part C:Emerging Technologies, 2019, 101:297-312.
[11]胡永恺.基于手机信令的轨道交通乘客出行行为分析方法研究[D].南京:东南大学, 2017.HU Yongkai. Urban rail transit passenger travel behavior analysis methods based on cellular data[D]. Nanjing:Southeast University, 2017.
[12]DENG Y, ZHAO P. The impact of new metro on travel behavior:panel analysis using mobile phone data[J].Transportation Research Part A:Policy and Practice,2022, 162:46-57.
[13]DUAN Z, LIU X, YU Q, et al. Analyzing detour behavior of metro passengers based on mobile phone data[J].Transportation Planning and Technology, 2022, 45(3):289-309.
[14]殷勇,鞠子奇,吴雨遥,等.国外轨道交通发展对我国城市群轨道交通一体化的启示[J].交通运输工程与信息学报, 2021, 19(1):52-58.YIN Yong, JU Ziqi, WU Yuyao, et al. The enlightenment of foreign rail transit development to the rail transit integration of China megalopolis[J]. Journal of Transportation Engineering and Information, 2021,19(1):52-58.
[15]刘梦琪,瞿何舟.基于轨道交通与常规公交组合的出行路径选择研究[J].交通运输工程与信息学报, 2018,16(4):63-68.LIU Mengqi, QU Hezhou. Study on the route choice using the combination of rail and bus transit[J]. Journal of Transportation Engineering and Information, 2018, 16(4):63-68.
[16]冉斌.手机数据在交通调查和交通规划中的应用[J].城市交通, 2013, 11(1):72-81, 32.RAN Bin. Use of cellphone data in travel survey and transportation planning[J]. Urban Transport of China,2013, 11(1):72-81, 32.
[17]李玮锋.基于移动通信数据的居民活动空间分析[D].上海：同济大学, 2018.LI Weifeng. Analysis on Individuals’Activity Space Based on Mobile Phone Data[D]. Shanghai:Tongji University,2018.
[18]WANG Y, DE ALMEIDA CORREIA G H, DE ROMPH E, et al. Using metro smart card data to model location choice of after-work activities:an application to Shanghai[J]. Journal of Transport Geography, 2017, 63:40-47.
[19]BENESTY J, CHEN J, HUANG Y, et al. Pearson correlation coefficient[M]//Noise Reduction in Speech Processing. Berlin, Heidelberg:Springer Berlin Heidelberg,2009:1-4.
[20]WANG L, CHEN Y, WANG Y, et al. Identification and classification of bus and subway passenger travel patterns in Beijing using transit smart card data[J]. Journal of Advanced Transportation, 2023:6529819.
[21]NISHIUCHI H, KING J, TODOROKI T. Spatial-temporal daily frequent trip pattern of public transport passengers using smart card data[J]. International Journal of Intelligent Transportation Systems Research, 2013, 11(1):1-10.
[22]ARTHUR D, VASSILVITSKII S. K-means++:the advantages of careful seeding[R]. Stanford, 2006.
[23]SHAHAPURE K R, NICHOLAS C. Cluster quality analysis using silhouette score[C]//2020 IEEE 7th International Conference on Data Science and Advanced Analytics(DSAA). Sydney, Australia:IEEE, 2020:747-748.
基本信息:
DOI：10.19961/j.cnki.1672-4747.2025.07.026
中图分类号:U293.6
引用信息:
[1]刘晓磊,邹国建,段征宇,等.面向个体出行的地铁路径提取与行为模式挖掘[J].交通运输工程与信息学报,2026,24(01):15-24.DOI:10.19961/j.cnki.1672-4747.2025.07.026.
基金信息:
上海市“科技创新行动计划”社会发展科技攻关项目（20dz1202903）