MvHSTM: A Multi-view Hypergraph Spatio-Temporal Model for Traffic Speed Forecasting
Keywords: Traffic Prediction, Deep Learning, Hypergraph Convolution
TL;DR: We propose a spatio-temporal model to address traffic speed forecasting, incorporating hypergraph convolution and a temporal transformer.
Abstract: Accurate traffic speed prediction is critical in modern society as it is effective for both individuals and authorities. Due to the large scale of urban road networks, traffic speed exhibits complex spatio-temporal dependencies, not only among adjacent nodes but also across the network, reflecting both local and cross-regional simultaneous correlations. However, existing studies have not effectively addressed these characteristics. In this context, we propose a novel framework called Multi-view Hypergraph Spatio-Temporal Model (MvHSTM) that employs a temporal transformer to capture temporal dependencies and utilizes hypergraph convolutional networks to inherently model spatial relationships. Specifically, we introduce two hypergraph construction methods, the Geographical Adjacency Hypergraph (GAH) and the Feature Similarity Hypergraph (FSH), to capture spatial correlations on neighboring and non-neighboring scales. Extensive experiments on real-world traffic speed datasets demonstrate that our approach achieves state-of-the-art performance compared to baseline methods.
Supplementary Material: zip
Primary Area: other topics in machine learning (i.e., none of the above)
Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics.
Submission Guidelines: I certify that this submission complies with the submission instructions as described on https://iclr.cc/Conferences/2025/AuthorGuide.
Reciprocal Reviewing: I understand the reciprocal reviewing requirement as described on https://iclr.cc/Conferences/2025/CallForPapers. If none of the authors are registered as a reviewer, it may result in a desk rejection at the discretion of the program chairs. To request an exception, please complete this form at https://forms.gle/Huojr6VjkFxiQsUp6.
Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.
No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.
Submission Number: 13669
Loading