Abstract:Real-time and precise traffic flow prediction is vital for the efficiency of intelligent transportation systems. Traditional methods often employ graph neural networks (GNNs) with predefined graphs to describe spatial correlations among traffic nodes in urban road networks. However, these pre-defined graphs are limited by existing knowledge and graph generation methodologies, offering an incomplete picture of spatial correlations. While time-varying graphs based on data-driven learning have attempted to address these limitations, they still struggle with adequately capturing the inherent spatial correlations in traffic data. Moreover, most current methods for capturing dynamic temporal correlations rely on a unified calculation scheme using a temporal multi-head self-attention mechanism, which at some level might leads to inaccuracies. In order to overcome these challenges, we have proposed a novel hybrid time-varying graph neural network (HTVGNN) for traffic flow prediction. Firstly, a novel enhanced temporal perception multi-head self-attention mechanism based on time-varying mask enhancement was reported to more accurately model the dynamic temporal dependencies among distinct traffic nodes in the traffic network. Secondly, we have proposed a novel graph learning strategy to concurrently learn both static and dynamic spatial associations between different traffic nodes in road networks. Meanwhile, in order to enhance the learning ability of time-varying graphs, a coupled graph learning mechanism was designed to couple the graphs learned at each time step. Finally, the effectiveness of the proposed method HTVGNN was demonstrated with four real data sets. Simulation results revealed that HTVGNN achieves superior prediction accuracy compared to the state of the art spatio-temporal graph neural network models. Additionally, the ablation experiment verifies that the coupled graph learning mechanism can effectively improve the long-term prediction performance of HTVGNN.

What problem does this paper attempt to address?

### Problems the paper attempts to solve This paper aims to solve the problem of real - time and accurate traffic flow prediction in intelligent transportation systems. Specifically, the article points out the following limitations of current traffic flow prediction methods: 1. **Limitations of pre - defined graph structures**: - Traditional methods usually use pre - defined graph neural networks (GNNs) to describe the spatial correlations between different traffic nodes in urban road networks. However, these pre - defined graphs rely on existing knowledge and graph generation methods and cannot comprehensively and accurately represent spatial correlations. 2. **Insufficient capture of dynamic spatio - temporal correlations**: - Although time - varying graphs based on data - driven learning attempt to solve the above problems, they still have deficiencies in capturing the inherent spatial correlations in traffic data. - Most existing methods rely on a unified temporal multi - head self - attention mechanism, which may lead to a decrease in accuracy. 3. **Limitations of static spatial correlations**: - Existing models such as STGHTN and CorrSTN have made some improvements in dynamic modeling, but their computational complexity is high, and they fail to fully consider the coupling relationships between time - varying graphs at different time steps. To solve these problems, the authors propose a new Hybrid Time - Varying Graph Neural Network (HTVGNN). The main innovations of this model include: 1. **Enhanced temporal perception multi - head self - attention mechanism**: - An Enhanced Temporal Perception Multi - Head Self - Attention Mechanism based on time - varying mask enhancement is proposed to more accurately model the dynamic temporal dependencies between different traffic nodes. 2. **Novel graph learning strategy**: - Simultaneously learn static and dynamic spatial associations, and propose a Coupled Graph Learning Mechanism to enhance the learning ability of time - varying graphs. 3. **Convolutional recurrent module combining static and dynamic graphs**: - Design a Coupled Time - Varying Graph Convolution Gated Recurrent Module (CTVGCRM) to extract spatio - temporal correlations. Through these improvements, HTVGNN can achieve higher prediction accuracy than the existing state - of - the - art spatio - temporal graph neural network models on four real - world datasets. In addition, ablation experiments verify the effective improvement of the coupled graph learning mechanism on long - term prediction performance. ### Summary By introducing an enhanced temporal perception multi - head self - attention mechanism, a novel graph learning strategy, and a coupled time - varying graph convolution recurrent module, this paper solves the deficiencies of existing traffic flow prediction methods in capturing dynamic spatio - temporal correlations and spatial associations, thereby improving the prediction accuracy.

A novel hybrid time-varying graph neural network for traffic flow forecasting

TF-MVGNN: an accurate traffic forecasting framework based on spatial–temporal graph neural network through exploiting multiple-view graph construction and learning

A Novel Attention-Based Dynamic Multi-Graph Spatial-Temporal Graph Neural Network Model for Traffic Prediction

Traffic Flow Prediction via Spatial Temporal Graph Neural Network

Learning Dynamics and Heterogeneity of Spatial-Temporal Graph Data for Traffic Forecasting

Temporal Graph Convolutional Network for Urban Traffic Flow Prediction Method.

Spatial Temporal Incidence Dynamic Graph Neural Networks for Traffic Flow Forecasting

A Spatio-Temporal Graph Neural Network Approach for Traffic Flow Prediction

Adaptive Hybrid Spatial-Temporal Graph Neural Network for Cellular Traffic Prediction

Adaptive spatial-temporal graph attention networks for traffic flow forecasting

Dynamic Hypergraph Structure Learning for Traffic Flow Forecasting

Dynamic Spatial–temporal Graph Convolutional Recurrent Networks for Traffic Flow Forecasting

Adaptive Graph Generation Based on Generalized Pagerank Graph Neural Network for Traffic Flow Forecasting

Hybrid Spatio-Temporal Graph Convolution Network for Short-Term Traffic Forecasting.

Geometric Algebra Multi-order Graph Neural Network for Traffic Prediction

Temporal Graph Learning Recurrent Neural Network for Traffic Forecasting

T-GCN: A Temporal Graph ConvolutionalNetwork for Traffic Prediction

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

Spatial-temporal multi-factor fusion graph neural network for traffic prediction

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

Graph learning-based spatial-temporal graph convolutional neural networks for traffic forecasting