🤖 AI Summary
To address the challenge of accurate traffic flow forecasting in data-scarce road networks, this paper proposes a Transfer Learning-enhanced Graph Pruning Spatio-Temporal Graph Convolutional Network (TL-GPSTGN). The method integrates adaptive graph pruning with transfer learning: first, an entropy-driven pruning mechanism—guided by edge correlation and information entropy—dynamically simplifies the adjacency matrix to enhance structural generalizability and cross-domain transferability; second, an end-to-end transfer learning framework is designed to mitigate overfitting under limited training samples. Extensive experiments on real-world traffic datasets demonstrate that TL-GPSTGN achieves a 12.7% reduction in mean absolute error (MAE) for single-dataset prediction and an average 19.3% decrease in root mean square error (RMSE) for cross-city transfer tasks. These results significantly outperform state-of-the-art spatio-temporal graph convolutional models (e.g., ST-GCN variants), validating both its robustness in low-data regimes and its effectiveness in knowledge transfer across heterogeneous urban networks.
📝 Abstract
With the process of urbanization and the rapid growth of population, the issue of traffic congestion has become an increasingly critical concern. Intelligent transportation systems heavily rely on real-time and precise prediction algorithms to address this problem. While Recurrent Neural Network (RNN) and Graph Convolutional Network (GCN) methods in deep learning have demonstrated high accuracy in predicting road conditions when sufficient data is available, forecasting in road networks with limited data remains a challenging task. This study proposed a novel Spatial-temporal Convolutional Network (TL-GPSTGN) based on graph pruning and transfer learning framework to tackle this issue. Firstly, the essential structure and information of the graph are extracted by analyzing the correlation and information entropy of the road network structure and feature data. By utilizing graph pruning techniques, the adjacency matrix of the graph and the input feature data are processed, resulting in a significant improvement in the model's migration performance. Subsequently, the well-characterized data are inputted into the spatial-temporal graph convolutional network to capture the spatial-temporal relationships and make predictions regarding the road conditions. Furthermore, this study conducts comprehensive testing and validation of the TL-GPSTGN method on real datasets, comparing its prediction performance against other commonly used models under identical conditions. The results demonstrate the exceptional predictive accuracy of TL-GPSTGN on a single dataset, as well as its robust migration performance across different datasets.