UNIFYING LONG AND SHORT SPATIO-TEMPORAL FORECASTING WITH SPECTRAL GRAPH NEURAL NETWORKS
Keywords: multivariate time series forecasting, spatio-temporal graph neural network, spectral graph neural network
TL;DR: Leveraging the ability of spectral GNN to capture global features, we propose an efficient STGNN model capable of handling long sequences.
Abstract: Multivariate Time Series (MTS) forecasting plays a vital role in various practical applications. Current research in this area is categorized into Spatial-Temporal Forecasting (STF) and Long-term Time Series Forecasting (LTSF). While these tasks share similarities, the methods and benchmarks used differ significantly. Spatio-Temporal Graph Neural Networks (STGNNs) excel at modeling interrelationships in STF tasks but face difficulties with long sequence inputs due to inefficient training. In contrast, LTSF models handle long sequences well but struggle with capturing complex variable interrelationships. This paper proposes the Spectral Spatio-Temporal Graph Neural Network (S2GNN) to address these challenges, unifying short- and long-sequence spatio-temporal forecasting within a single framework. S2GNN leverages spectral GNNs for global feature extraction incorporates an adaptive graph structure to manage varying sequence lengths and adopts a decoupled framework to improve scalability. Additionally, we introduce scale-adaptive node embeddings and cross-correlation embeddings for better differentiation between similar temporal patterns. Extensive experiments on eight public datasets, including both STF and LTSF datasets, demonstrate that S2GNN consistently outperforms state-of-the-art models across diverse prediction tasks. Code is available at \url{https://anonymous.4open.science/r/S2GNN-B21D}.
Primary Area: learning on time series and dynamical systems
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.
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: 9879
Loading