A Deep Learning Based Approach to Examine the impact of Land Use Change on Shared Bicycles Usage in Rail Transit Stations

• Published in: Tehnički vjesnik Vol. 32; no. 3; pp. 1113 - 1124
• Main Author: Chen, Xiang
• Format: Journal Article Paper
• Language: English
• Published: Slavonski Baod Josipa Jurja Strossmayer University of Osijek 2025; Sveučilište u Slavonskom Brodu, Stojarski fakultet; Faculty of Mechanical Engineering in Slavonski Brod, Faculty of Electrical Engineering in Osijek, Faculty of Civil Engineering in Osijek
• Subjects: Analytic hierarchy process; Bicycles; Cluster analysis; Deep learning; Land use; land use change; rail transit station; Rail transportation; shared bicycle; Traffic congestion; Urban areas
• ISSN: 1330-3651; 1848-6339
• DOI: 10.17559/TV-20240322001422

Due to China's booming economy and rapid urban expansion, significant progress has been made in the construction of rail transit (hereinafter referred to as RT), which is becoming an important means of connecting people in different urban areas, relieving urban traffic congestion and reducing environmental pollution. At the same time, the construction of rail transit will also affect the land use of the areas along the route, especially the mix of land use around the rail transit stations, and the mix of land use functions around the rail transit stations will affect the composition of the population around the stations, which in turn will affect the choice of feeder transport for passengers leaving the stations. From the perspective of environmental friendliness, low carbon and economic convenience, bicycle sharing is the best option to solve the problem of "last mile" to the destination after leaving the station, and its usage rate is the highest.Therefore, this study establishes a deep learning technology method to study the possibility of the impact of land use on shared bicycle usage within the influence range of rail transit stations. Firstly, the time-varying patronage of RT stations was predicted using cluster analysis and random forest method, then an evaluation system of the efficiency of shared units and RT interchange was constructed using AHP hierarchical analysis, followed by a vector machine model supported by least squares to study the relationship between land use and shared unit usage around RT stations. The results show that it is feasible to use deep learning techniques to study the impact of land use on bike-sharing usage within the area of influence of rail transit stations, and the case validation shows that an increase of 0.12 in the land use mix index at station A results in a 10.43% increase in share of shared bicycles in model choice, and an increase of 0.09 in the land use mix index at station B results in a 13.11% increase in share of shared bicycles in model choice. This indicates that with the development of rail transport and the increase in land use mix in the surrounding area, the share of shared bicycles in model choice is also increasing.