Research Article
Dynamic Traffic Network Based Multi-Modal Travel Mode Fusion Recommendation
@INPROCEEDINGS{10.1007/978-3-030-92635-9_18, author={Nannan Jia and Mengmeng Chang and Zhiming Ding and Zunhao Liu and Bowen Yang and Lei Yuan and Lutong Li}, title={Dynamic Traffic Network Based Multi-Modal Travel Mode Fusion Recommendation}, proceedings={Collaborative Computing: Networking, Applications and Worksharing. 17th EAI International Conference, CollaborateCom 2021, Virtual Event, October 16-18, 2021, Proceedings, Part I}, proceedings_a={COLLABORATECOM}, year={2022}, month={1}, keywords={Heterogeneous information network Graph neural network Meta-path Multi-modal travel mode Fusion recommendation}, doi={10.1007/978-3-030-92635-9_18} }- Nannan Jia
Mengmeng Chang
Zhiming Ding
Zunhao Liu
Bowen Yang
Lei Yuan
Lutong Li
Year: 2022
Dynamic Traffic Network Based Multi-Modal Travel Mode Fusion Recommendation
COLLABORATECOM
Springer
DOI: 10.1007/978-3-030-92635-9_18
Abstract
The travel problem is a challenge to the city’s overall development. With the increase of the number of cars and the increase of urban population density, intelligent travel mode provides new solutions to solve these problems. However, the existing research on the choice of travel modes for residents only considers the current traffic conditions, and the preferences of individual users for travel modes are poorly considered, which cannot meet the personalized travel needs of users. From this perspective, a heterogeneous information network based on users’ spatial-temporal travel trajectories is proposed in this paper. Considering the dynamic traffic network that is constantly changing during the travel process, and using the graph neural network guided by the meta-path to dynamically model the user and travel mode. Features embedding with rich interactive information, so as to fully learn the users’ preferences for travel modes in the time-space travel trajectory, and recommend travel modes that meet personalized needs to users. Finally, the effectiveness of the proposed method is demonstrated by experimental evaluation on real-world datasets.
