Abstract:In this work, we propose a deep neural network approach for traffic prediction, namely DSTN, which captures the dynamic nature in both temporal and spatial dimensions and periodicity of traffic data. In dynamic spatial modeling, unlike most existing attention‐based approaches, which generally model spatial dynamics by constructing a dynamic Laplace matrix based on an attention mechanism in a GCN or by using only an attention mechanism. We propose a joint latent space representation network for capturing spatial dynamics, which integrates GCN and attention mechanism in a novel way to compensate for the shortcomings of single information modeling and can capture global spatial dynamics adaptively. Multiple experiments on two real traffic datasets show that DSTN is superior to SOTA baseline in both road traffic and urban rail transit with good generalization capability. In the era of data‐driven transportation development, traffic forecasting is crucial. Established studies either ignore the inherent spatial structure of the traffic network or ignore the global spatial correlation and may not capture the spatial relationships adequately. In this work, a Dynamic Spatial‐Temporal Network (DSTN) based on Joint Latent Space Representation (JLSR) is proposed for traffic forecasting. Specifically, in the spatial dimension, a JLSR network is developed by integrating graph convolution and spatial attention operations to model complex spatial dependencies. Since it can adaptively fuse the representation information of local topological space and global dynamic space, a more comprehensive spatial dependency can be captured. In the temporal dimension, a Stacked Bidirectional Unidirectional Gated Recurrent Unit (SBUGRU) network is developed, which captures long‐term temporal dependencies through both forward and backward computations and superimposed recurrent layers. On these bases, DSTN is developed in an encoder‐decoder framework and periodicity is flexibly modeled by embedding branches. The performance of DSTN is validated on two types of real‐world traffic flow datasets, and it improves over baselines.

JointSTNet: Joint Pre-Training for Spatial-Temporal Traffic Forecasting

Spatial–Temporal Deep Tensor Neural Networks for Large-Scale Urban Network Speed Prediction

Spatio-Temporal Joint Graph Convolutional Networks for Traffic Forecasting

Dynamic spatial‐temporal network for traffic forecasting based on joint latent space representation

PreSTNet: Pre-trained Spatio-Temporal Network for traffic forecasting

STJLA: A Multi-Context Aware Spatio-Temporal Joint Linear Attention Network for Traffic Forecasting

STSD: Modeling Spatial Temporal Staticity and Dynamicity in Traffic Forecasting

Deep Spatio-temporal Adaptive 3D Convolutional Neural Networks for Traffic Flow Prediction

Space Meets Time: Local Spacetime Neural Network For Traffic Flow Forecasting

Adaptive spatial-temporal graph attention networks for traffic flow forecasting

Unified Spatial-Temporal Neighbor Attention Network for Dynamic Traffic Prediction

STG4Traffic: A Survey and Benchmark of Spatial-Temporal Graph Neural Networks for Traffic Prediction

Reinforced Spatiotemporal Attentive Graph Neural Networks for Traffic Forecasting.

Forecast Network-Wide Traffic States for Multiple Steps Ahead: A Deep Learning Approach Considering Dynamic Non-Local Spatial Correlation and Non-Stationary Temporal Dependency

Dynamic multi-scale spatial-temporal graph convolutional network for traffic flow prediction

Adaptive and Interactive Multi-Level Spatio-Temporal Network for Traffic Forecasting

Ada-STNet: A Dynamic AdaBoost Spatio-Temporal Network for Traffic Flow Prediction.

Spatial–Temporal Dynamic Graph Convolutional Network With Interactive Learning for Traffic Forecasting

Combining jumping knowledge into traffic forecasting: An attention-based spatial-temporal adaptive integration gated network

Spatio-Temporal Graph Neural Network for Traffic Prediction Based on Adaptive Neighborhood Selection