面向动态交通流量预测的自适应图注意Transformer

刘宇轩,刘毅志,廖祝华,邹正标,汤璟昕

Adaptive graph attention Transformer for dynamic traffic flow prediction

Yuxuan LIU,Yizhi LIU,Zhuhua LIAO,Zhengbiao ZOU,Jingxin TANG

表 2 PEMS04数据集上不同模型的RMSE和MAE对比结果

Tab.2 Comparison results of RMSE and MAE for different models on PEMS04 dataset

模型 Tp=15 min Tp=30 min Tp=45 min Tp=60 min RMSE MAE RMSE MAE RMSE MAE RMSE MAE STGCN 30.49 19.98 33.23 21.46 36.87 24.48 39.41 26.93 T-GCN 29.39 20.32 32.79 21.94 36.27 24.06 39.88 27.27 DCRNN 28.65 19.06 32.72 22.09 35.74 23.89 41.19 28.51 DMSTGCN 28.02 18.81 31.46 21.02 34.38 22.64 38.67 26.15 Trendformer 27.47 18.62 30.59 20.56 33.53 22.32 34.17 23.54 STTN 26.95 17.98 29.76 19.48 31.28 21.76 33.35 22.43 ATST-GCN 27.32 18.45 30.19 20.07 30.24 21.03 32.49 22.17 MSAGAFormer 25.78 16.33 27.51 17.42 27.13 17.94 28.89 18.76