🤖 AI Summary
Traffic flow forecasting faces dual challenges: difficulty in modeling spatiotemporal coupling and low computational efficiency—existing methods either decouple spatial and temporal dependencies or jointly model them at the cost of accuracy and speed. This paper proposes STEI-PCN, a purely convolutional architecture featuring a novel dynamic graph structure inference module grounded in absolute and relative spatiotemporal coordinates, enabling data-driven, adaptive adjacency matrix learning. It integrates local synchronous spatiotemporal convolutions with long-range temporal dilated causal convolutions to jointly capture fine-grained spatiotemporal interactions and long-term dependencies. Additionally, a multi-view gated collaborative prediction mechanism enhances robustness. STEI-PCN achieves state-of-the-art (SOTA) or near-SOTA performance on PeMS03/04/07/08 and PeMS-Bay, while significantly outperforming mainstream spatiotemporal models in both training and inference speed.
📝 Abstract
Traffic data exhibits complex temporal, spatial, and spatial-temporal correlations. Most of models use either independent modules to separately extract temporal and spatial correlations or joint modules to synchronously extract them, without considering the spatial-temporal correlations. Moreover, models that consider joint spatial-temporal correlations (temporal, spatial, and spatial-temporal correlations) often encounter significant challenges in accuracy and computational efficiency which prevent such models from demonstrating the expected advantages of a joint spatial-temporal correlations architecture. To address these issues, this paper proposes an efficient pure convolutional network for traffic prediction via spatial-temporal encoding and inferring (STEI-PCN). The model introduces and designs a dynamic adjacency matrix inferring module based on absolute spatial and temporal coordinates, as well as relative spatial and temporal distance encoding, using a graph convolutional network combined with gating mechanism to capture local synchronous joint spatial-temporal correlations. Additionally, three layers of temporal dilated causal convolutional network are used to capture long-range temporal correlations. Finally, through multi-view collaborative prediction module, the model integrates the gated-activated original, local synchronous joint spatial-temporal, and long-range temporal features to achieve comprehensive prediction. This study conducts extensive experiments on flow datasets (PeMS03/04/07/08) and speed dataset (PeMS-Bay), covering multiple prediction horizons. The results show that STEI-PCN demonstrates competitive computational efficiency in both training and inference speeds, and achieves superior or slightly inferior to state-of-the-art (SOTA) models on most evaluation metrics.
