Answers are grounded in this paper and cite their sources.
Urban transportation systems are increasingly dependent on accurate demand forecasting to ensure efficient resource allocation and service optimization. Traditional methods such as ARIMA and LSTM often fail to capture complex spatial–temporal dependencies and cyclical patterns, leading to limited predictive accuracy. To address these challenges, this study proposes a novel spatiotemporal convolutional neural network enhanced with attention and Fourier transform (ST‐CNN‐AF). The model integrates CNN for extracting spatial features, self‐attention for capturing long‐range temporal dependencies, and Fourier transform to effectively model periodic demand fluctuations. Extensive experiments were conducted on real‐world taxi demand datasets, and the proposed model achieved significant improvements, reducing RMSE by 6.6% and MAE by 4.7% compared to baseline models. These results demonstrate the superior ability of ST‐CNN‐AF to capture complex urban mobility patterns. Beyond performance gains, the model provides a practical tool for intelligent transportation systems, enabling applications in taxi dispatching, traffic management, and dynamic pricing strategies. This work contributes to bridging the gap between deep learning advancements and real‐world transportation needs, offering both methodological innovation and practical value.
No takes yet. Share an insight, caveat, or question.
A 2025 study studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: