Metro Station Area Functional Classification Based on Graph Convolutional Networks and Integrated Metro Passenger Flow Data
Year: 2025
Type: article
Abstract: ABSTRACT The analysis of diverse human mobility data plays a significant role in the revelation of the functional structure of cities. The growing popularity of metro travel in recent years has afforded new opportunities for examining the functional structure of cities. However, there has been inadequate exploration of the traffic interaction patterns embedded within metro passenger flow data, which refers to the patterns of movement and interaction between metro stations based on passenger flow information. Existing studies often overlook the potential of these interaction data for identifying the regional functions of metro stations. To address this issue, this study proposes a geosemantic analysis framework for modeling the relationship between traffic spatial interaction patterns and functional structures at the metro station level. Firstly, a corpus of metro station trajectories is constructed using metro passenger flow data, and high‐dimensional semantic embedding techniques are employed to extract feature representations of the metro stations. Secondly, in consideration of the interconnected nature of metro stations within the broader metro network, this study employs a graph convolutional neural network (GCNN) to analyze contextual and network topology data in order to predict the functional classification of metro stations. The efficacy of the proposed framework is validated through an empirical investigation conducted in Chongqing Municipality. In comparison to existing methodologies, this study's approach demonstrates superior performance in the functional classification of metro station areas, with an average accuracy of 0.894. The findings of this study facilitate a deeper comprehension of the interrelationship between urban transportation modes and functional structures, and provide a scientific foundation and a practical framework for urban planners and transportation managers in the design of strategies for the interaction between transportation and urban structures.
Source: Transactions in GIS
DOI: 10.1111/tgis.70151
Authors: Liuchang Xu, Huanbing Ren, Ruichen Mao, Junwei Yao, Chengkun Zhang, Sensen Wu, Hailin Feng, Zhenhong Du
Institutions: Zhejiang A & F University, Zhejiang University
Language: English
Subfield: Transportation
Field: Social Sciences
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W4416308998