近年來，隨著人們對交通運輸工具的依賴性提高，制訂完善的交通管理方針成為眾所關注的議題。交通相關的預測在決定交通政策方面起著至關重要的作用；傳統的方法只能根據統計結果或歷史經驗做出決定。而本項研究通過機器學習，我們能夠捕捉到城市動態之間的潛在相互作用，並在空間範圍內找到它們的時空相關性。然而，儘管過去有大量的交通相關研究工作，但其中卻很少有研究探討預測擁堵的影響。因此，本論文探討的重點是在高速公路上預測車禍導致交通擁堵的影響程度，特別是關於擁堵發生的時間長度。因此，我們提出了一個名為雙注意多尺度圖卷積網絡（DAMGNet）的新型模型來解決這個問題。在這個提議的模型中，異質數據，如事故訊息、城市動態和高速公路網絡的各種特徵，被AI考慮並結合起來。我們利用上下文編碼器對事故數據進行編碼，並使用空間編碼器捕捉多尺度圖卷積網絡（GCN）之間的隱藏特徵。通過我們設計的雙重關注機制，DAMGNet模型能夠有效地學習特徵之間的關聯性。我們使用真實世界的數據集上進行的評估證明， DAMGNet在RMSE和MAE等預測數值方面比其他比較方法有明顯的改善。我們的模型可以在未來的交通決策上，提供管理者更加精準的判別依據，進而提升用路的道路使用品質。

Traffic-related forecasting plays a critical role in determining transportation policy; unlike traditional approaches, which can only make decisions based on statistical results or historical experience. Through machine learning, we are able to capture the potential interactions between urban dynamics and find their mutual interactions in a spatial context. However, despite a plethora of traffic-related studies, few works have explored predicting the impact of congestion. Therefore, this paper focuses on predicting how a car accident leads to traffic congestion, especially about the length of time it takes for the congestion to occur. Accordingly, we propose a novel model named Dual-Attention Multi-Scale Graph Convolutional Networks (DAMGNet) to address this issue. In this proposed model, heterogeneous data such as accident information, urban dynamics and various characteristics for highway networks, are considered and combined. Next, the context encoder encodes the accident data and the spatial encoder captures the hidden features between multi-scale Graph Convolutional Networks (GCNs). With our designed dual attention mechanism, the DAMGNet model is able to effectively learn the correlation between features. The evaluations conducted on a real-world dataset prove that our DAMGNet has a significant improvement in RMSE and MAE over other comparative methods.

[1] Baykal-Gursoy, M., Xiao, W., Duan, Z. and Ozbay, K. Delay estimation for traffic flow interrupted by incidents. In 86th Annual Transportation Research Conf., Transportation Research Board, Washington, DC (2006)2006.
[2] Breiman, L. Random forests. Machine learning, 45, 1 (2001), 5-32.
[3] Bruna, J., Zaremba, W., Szlam, A. and LeCun, Y. Spectral networks and locally connected networks on graphs. arXiv preprint arXiv:1312.6203 (2013).
[4] Chen, T. and Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining (2016)2016.
[5] Chien, S., Yang, Z. and Hou, E. Genetic algorithm approach for transit route planning and design. Journal of transportation engineering, 127, 3 (2001), 200-207.
[6] Defferrard, M., Bresson, X. and Vandergheynst, P. Convolutional neural networks on graphs with fast localized spectral filtering. arXiv preprint arXiv:1606.09375 (2016).
[7] Ge, L., Li, S., Wang, Y., Chang, F. and Wu, K. Global Spatial-Temporal Graph Convolutional Network for Urban Traffic Speed Prediction. Applied Sciences, 10, 4 (2020), 1509.
[8] Grover, A. and Leskovec, J. node2vec: Scalable feature learning for networks. In Proceedings of the 22nd ACM SIGKDD international conference on Knowledge discovery and data mining (2016)2016.
[9] Guo, C. and Berkhahn, F. Entity embeddings of categorical variables. arXiv preprint arXiv:1604.06737 (2016).
[10] Guo, S., Lin, Y., Feng, N., Song, C. and Wan, H. Attention based spatial-temporal graph convolutional networks for traffic flow forecasting. In Proceedings of the AAAI Conference on Artificial Intelligence (2019)2019.
[11] Hamilton, W. L., Ying, R. and Leskovec, J. Representation learning on graphs: Methods and applications. arXiv preprint arXiv:1709.05584 (2017).
[12] Huang, R., Huang, C., Liu, Y., Dai, G. and Kong, W. LSGCN: Long Short-Term Traffic Prediction with Graph Convolutional Networks. IJCAI (2020).
[13] Ioffe, S. and Szegedy, C. Batch normalization: Accelerating deep network training by reducing internal covariate shift. In International conference on machine learning (2015). PMLR, 2015.
[14] Kipf, T. N. and Welling, M. Semi-supervised classification with graph convolutional networks. arXiv preprint arXiv:1609.02907 (2016).
[15] LeCun, Y., Bengio, Y. and Hinton, G. Deep learning. nature, 521, 7553 (2015), 436-444.
[16] Li, Y., Yu, R., Shahabi, C. and Liu, Y. Diffusion convolutional recurrent neural network: Data-driven traffic forecasting. arXiv preprint arXiv:1707.01926 (2017).
[17] Liu, D., Li, J., Du, B., Chang, J. and Gao, R. Daml: Dual attention mutual learning between ratings and reviews for item recommendation. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (2019)2019.
[18] Liu, D., Tang, L., Shen, G. and Han, X. Traffic speed prediction: an attention-based method. Sensors, 19, 18 (2019), 3836.
[19] Liu, X., Chien, S. I. and Chen, M. An adaptive model for highway travel time prediction. Journal of Advanced Transportation, 48, 6 (2014), 642-654.
[20] Lu, B., Gan, X., Jin, H., Fu, L. and Zhang, H. Spatiotemporal Adaptive Gated Graph Convolution Network for Urban Traffic Flow Forecasting. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management (2020)2020.
[21] Lv, Z., Xu, J., Zheng, K., Yin, H., Zhao, P. and Zhou, X. Lc-rnn: A deep learning model for traffic speed prediction. In IJCAI (2018)2018.
[22] MacKay, M. Traffic accidents—a modern epidemic. International Journal of Environmental Studies, 3, 1-4 (1972), 223-227.
[23] Mangharam, R., Lee, I. and Sokolsky, O. Real-Time Traffic Congestion Prediction (2008).
[24] Mátrai, T., Tóth, J. and Horváth, M. T. Route planning based on urban mobility management. Hungarian Journal of Industry and Chemistry (2016), 71-79.
[25] Priambodo, B. and Jumaryadi, Y. Time series traffic speed prediction using k-nearest neighbour based on similar traffic data. In MATEC Web of Conferences (2018). EDP Sciences, 2018.
[26] Ramachandran, P., Parmar, N., Vaswani, A., Bello, I., Levskaya, A. and Shlens, J. Stand-alone self-attention in vision models. arXiv preprint arXiv:1906.05909 (2019).
[27] Škrlj, B., Džeroski, S., Lavrač, N. and Petkovič, M. Feature importance estimation with self-attention networks. arXiv preprint arXiv:2002.04464 (2020).
[28] Szmelter, A. The importance of automotive industry in shaping habitants mobility in future cities. Transport Economics and Logistics, 71 (2017), 163-178.
[29] Vashishth, S., Yadav, P., Bhandari, M. and Talukdar, P. Confidence-based graph convolutional networks for semi-supervised learning. In The 22nd International Conference on Artificial Intelligence and Statistics (2019). PMLR, 2019.
[30] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L. and Polosukhin, I. Attention is all you need. arXiv preprint arXiv:1706.03762 (2017).
[31] Veličković, P., Cucurull, G., Casanova, A., Romero, A., Lio, P. and Bengio, Y. Graph attention networks. arXiv preprint arXiv:1710.10903 (2017).
[32] Wu, Z., Pan, S., Long, G., Jiang, J. and Zhang, C. Graph wavenet for deep spatial-temporal graph modeling. arXiv preprint arXiv:1906.00121 (2019).
[33] Xie, Q., Guo, T., Chen, Y., Xiao, Y., Wang, X. and Zhao, B. Y. Deep Graph Convolutional Networks for Incident-Driven Traffic Speed Prediction. In Proceedings of the 29th ACM International Conference on Information & Knowledge Management (2020)2020.
[34] Yassin, S. S. Road accident prediction and model interpretation using a hybrid K-means and random forest algorithm approach. SN Applied Sciences, 2, 9 (2020), 1-13.
[35] Ye, S. Research on urban road traffic congestion charging based on sustainable development. Physics Procedia, 24 (2012), 1567-1572.
[36] Yu, B., Yin, H. and Zhu, Z. Spatio-temporal graph convolutional networks: A deep learning framework for traffic forecasting. arXiv preprint arXiv:1709.04875 (2017).
[37] Yuan, Z., Zhou, X. and Yang, T. Hetero-convlstm: A deep learning approach to traffic accident prediction on heterogeneous spatio-temporal data. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (2018)2018.
[38] Zafar, N. and Ul Haq, I. Traffic congestion prediction based on Estimated Time of Arrival. PloS one, 15, 12 (2020), e0238200.
[39] Zhang, Z., Yang, W. and Wushour, S. Traffic accident prediction based on LSTM-GBRT model. Journal of Control Science and Engineering, 2020 (2020).
[40] Zhao, H., Cheng, H., Mao, T. and He, C. Research on Traffic Accident Prediction Model Based on Convolutional Neural Networks in VANET. In 2019 2nd International Conference on Artificial Intelligence and Big Data (ICAIBD) (2019). IEEE, 2019.
[41] Zhao, J., Gao, Y., Tang, J., Zhu, L. and Ma, J. Highway travel time prediction using sparse tensor completion tactics and-nearest neighbor pattern matching method. Journal of Advanced Transportation, 2018 (2018).
[42] Zheng, C., Fan, X., Wang, C. and Qi, J. Gman: A graph multi-attention network for traffic prediction. In Proceedings of the AAAI Conference on Artificial Intelligence (2020)2020.