Spatiotemporal-aware Trend-Seasonality Decomposition Network for Traffic Flow Forecasting

TL;DR

This paper introduces STDN, a novel spatiotemporal model for traffic flow forecasting that effectively captures complex dynamics and decomposes traffic signals into trend and seasonal components, leading to improved accuracy.

Contribution

The paper proposes a new spatiotemporal-aware network with a trend-seasonality decomposition module and a dynamic graph structure, advancing traffic prediction methods.

Findings

STDN outperforms existing models on real-world datasets.

The model achieves high accuracy with low computational cost.

The new JiNan dataset enriches traffic prediction evaluation scenarios.

Abstract

Traffic prediction is critical for optimizing travel scheduling and enhancing public safety, yet the complex spatial and temporal dynamics within traffic data present significant challenges for accurate forecasting. In this paper, we introduce a novel model, the Spatiotemporal-aware Trend-Seasonality Decomposition Network (STDN). This model begins by constructing a dynamic graph structure to represent traffic flow and incorporates novel spatio-temporal embeddings to jointly capture global traffic dynamics. The representations learned are further refined by a specially designed trend-seasonality decomposition module, which disentangles the trend-cyclical component and seasonal component for each traffic node at different times within the graph. These components are subsequently processed through an encoder-decoder network to generate the final predictions. Extensive experiments conducted on real-world traffic datasets demonstrate that STDN achieves superior performance with remarkable computation cost. Furthermore, we have released a new traffic dataset named JiNan, which features unique inner-city dynamics, thereby enriching the scenario comprehensiveness in traffic prediction evaluation.