Table 1 The Computation cost on the PEMS04 dataset.
From: Multi-scale spatio-temporal graph neural network for urban traffic flow prediction
Method | Year | Publication | Highlight words |
|---|---|---|---|
STG-NCDE29 | 2022 | AAAI | GCNs, NCDEs, continuously changing spatio-temporal dependencies. |
ST-aware18 | 2022 | ICDE | Data-driven, spatio-temporal awareness, window attention mechanism |
DSTAGNN20 | 2022 | PMLR | Data-driven, dynamic spatio-temporal aware graph, graph convolution, gated convolution |
Auto-DSTSGN15 | 2023 | T-ITS | Data-driven, dilated convolution, spatio-temporal synchronous graph |
PDFormer30 | 2023 | AAAI | Spatio-temporal self-attention, transformer, heterogeneous attention fusion |
STGSA19 | 2023 | JAS | Data-driven, synchronous spatial-temporal graph, multi-head attention mechanism |
ISTNet16 | 2023 | DASFAA | Transformer, convolution network, maximum pooling, local and global dependencies |
DeepSTUQ7 | 2024 | TKDE | Data-driven, GCN, GRU, Uncertainty Quantification |
STJGCN17 | 2024 | TKDE | Dilated causal convolution, spatio-temporal joint graph, multi-range attention mechanism |
DEC-Former3 | 2024 | CIKM | Data-driven, spatio-temporal decoupling, attention mechanism |
STD-MAE31 | 2024 | IJCAI | Spatio-temporal decoupled masking, data pre-training, integrate downstream spatio-temporal prediction methods |
