| 76 | 0 | 63 |
| 下载次数 | 被引频次 | 阅读次数 |
本研究提出了一种基于变分模态分解(VMD)与自适应时空图神经网络(Ada-STNet)的交通流量预测模型，旨在综合处理数据的非平稳性和时空相关性，从而提高交通流量预测精度。首先，通过VMD将交通流量历史数据分解成若干平稳的本征模态函数(IMFs),消除非平稳性干扰提升后续建模效果。然后，针对每个IMFs分量输入到Ada-STNet模型，利用图卷积网络(GCN)动态捕捉交通路网的空间相关性，并通过门控时间卷积网络(TCN)提取时间依赖性。最终通过各分量结果重构获得完整预测值。实验结果表明，该模型在交通流量数据集上预测性能表现出色，和对比的次优模型比较，在平均绝对误差(MAE)、均方根误差(RMSE)和平均绝对百分比误差(MAPE)三项误差指标上分别降低了15.1%、28.7%和6.8%,验证了模型的有效性。
Abstract：A novel traffic flow prediction model based on variational mode decomposition(VMD) and adaptive spatiotemporal graph neural network(Ada-STNet) is proposed to address the non-stationarity and spatiotemporal correlation of data, thereby enhancing prediction accuracy. First, the VMD method decomposes historical traffic flow data into multiple stationary intrinsic mode functions(IMFs), mitigating the adverse effects of data non-stationarity and enhancing subsequent modeling performance. Each IMFs component is then processed by the Ada-STNet model, which employs graph convolutional network(GCN) to dynamically capture spatial correlations in the traffic network and gated temporal convolutional network(TCN) to extract temporal dependencies. Finally, the prediction results of each component are reconstructed to obtain the complete traffic flow prediction values. Experimental results demonstrate that the proposed model exhibits excellent prediction performance on a traffic flow dataset. Compared with the second-best model, it reduces the MAE, RMSE, and MAPE by 15.1%, 28.7%, and 6.8%, respectively, verifying its effectiveness.
[1] LU Saiqun,ZHANG Qiyan,CHEN Guangsen,et al.A combined method for short-term traffic flow prediction based on recurrent neural network[J].Alexandria Engineering Journal,2021,60(1):87-94.
[2] HSUEH Y L,YANG Yuren.A short-term traffic speed prediction model based on LSTM networks[J].International Journal of Intelligent Transportation Systems Research,2021,19(3):510-524.
[3] WU Pan,HUANG Zilin,PIAN Yuzhuang,et al.A combined deep learning method with attention-based LSTM model for short-term traffic speed forecasting[J].Journal of Advanced Transportation,2020,2020(1):1-15.
[4] 刘静，关伟.交通流预测方法综述[J].公路交通科技，2004,21(3):82-85.
[5] YE Qing,SZETO W Y,WONG S C.Short-term traffic speed forecasting based on data recorded at irregular intervals[J].IEEE Transactions on Intelligent Transportation Systems,2012,13(4):1727-1737.
[6] ZHANG Shuangyi,SONG Youyi,JIANG Dazhi,et al.Noise-identified kalman filter for short-term traffic flow forecasting[C]//2019 15th International Conference on Mobile Ad-Hoc and Sensor Networks (MSN).Shenzhen,China:IEEE,2019:462-466.
[7] 罗向龙，李丹阳，杨彧，等.基于KNN-LSTM的短时交通流预测[J].北京工业大学学报，2018,44(12):1521-1527.
[8] CORTES C,VAPNIK V.Support-vector networks[J].Machine Learning,1995,20(3):273-297.
[9] 李达，瞿伟，张勤，等.融合多层感知机和优化支持向量回归的滑坡位移预测模型[J].武汉大学学报(信息科学版),2023,48(8):1380-1388.
[10] ZAREMBA W,SUTSKEVER I,VINYALS O.Recurrent neural network regularization[DB/OL].(2014-09-08) [2025-05-15].https://arxiv.org/pdf/1409.2329v1.
[11] FENG Feifei,WU Xueling,NIU Ruiqing,et al.A landslide deformation analysis method using V/S and LSTM[J].Geomatics and Information Science of Wuhan University,2019,44(5):784-790.
[12] GUO Shengnan,LIN Youfang,LI Shijie,et al.Deep spatial-temporal 3D convolutional neural networks for traffic data forecasting[J].IEEE Transactions on Intelligent Transportation Systems,2019,20(10):3913-3926.
[13] LESOUPLE J,BAUDOIN C,SPIGAI M,et al.Generalized isolation forest for anomaly detection[J].Pattern Recognition Letters,2021,149:109-119.
[14] KIPF T N,WELLING M.Semi-supervised classification with graph convolutional networks[EB/OL].(2017-02-06) [2025-02-07].https://openreview.net/forum?id=SJU4ayYgl.
[15] 赵腾宇，李昕，黄晶晶.基于自适应时空图神经网络的交通预测[J].辽宁工业大学学报(自然科学版),2024,44(4):249-256.
[16] YU Bing,YIN Haoteng,ZHU Zhanxing.Spatio-temporal graph convolutional networks:a deep learning framework for traffic forecasting[C]//Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence.Stockholm,Sweden:AAAI Press,2018:3634-3640.
[17] TA Xuxiang,LIU Zihan,HU Xiao,et al.Adaptive spatio-temporal graph neural network for traffic forecasting[J].Knowledge-Based Systems,2022,242:108199.DOI:10.1016/j.knosys.2022.108199.
[18] DRAGOMIRETSKIY K,ZOSSO D.Variational mode decomposition[J].IEEE Transactions on Signal Processing,2014,62(3):531-544.
[19] 邱万锦.顾及时空特性和非平稳性的交通流深度预测方法[D].抚州：东华理工大学，2023.
[20] 龚循强，邱万锦，吕开云，等.变分模态分解与自适应图卷积门控循环网络的交通流量组合预测模型[J].武汉大学学报(信息科学版),2024,49(12):2329-2341.
[21] LI Fuxian,FENG Jie,YAN Huan,et al.Dynamic graph convolutional recurrent network for traffic prediction:benchmark and solution[J].ACM Transactions on Knowledge Discovery from Data,2023,17(1):1-21.
[22] LargeST(CCBY-TC 4.0)[EB/OL].[2025-01-23].https://github.com/liuxu77/LargeST?tab=readme-ov-file
[23] 黄河，谢军义，李志晖，等.基于门控时空图网络和TCN的交通流预测方法[J].交通科技，2024(6):126-131.
[24] HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[25] XIAO Hongxiang,ZHAO Zihan,YANG Tiejun.A traffic flow prediction method based on constrained dynamic graph convolutional recurrent networks[J].Engineering Applications of Artificial Intelligence,2024,133:108486.DOI:10.1016/j.engappai.2024.108486.
[26] LIU Xu,XIA Yutong,LIANG Yuxue,et al.Largest:a benchmark dataset for large-scale traffic forecasting[J].Advances in Neural Information Processing Systems,2023,36.75354-75371.
基本信息:
DOI：10.16191/j.cnki.hbkx.2025.06.002
中图分类号:TP183;U491.14
引用信息:
[1]韩珍珍,杨文焕,张嬿,等.基于变分模态分解与自适应时空图神经网络的交通流量预测[J].河北省科学院学报,2025,42(06):9-14+26.DOI:10.16191/j.cnki.hbkx.2025.06.002.
基金信息:
河北省科学院科技计划项目(25610); 石家庄市引进国外智力项目(20250015)
