Table 4 Path-aware models comparison
From: Gct-TTE: graph convolutional transformer for travel time estimation
Baseline/metric | Abakan | Omsk | ||||||
|---|---|---|---|---|---|---|---|---|
MAE | RMSE | MAPE | SR | MAE | RMSE | MAPE | SR | |
DeepIST | 153.88 | 241.29 | 0.3905 | 18.08 | 256.50 | 415.16 | 0.6361 | 14.39 |
DeepTTE | 111.03 | 174.56 | 0.2165 | 31.45 | 179.07 | 296.98 | 0.1898 | 34.03 |
GridLSTM | 100.27 | 206.91 | 0.2202 | 30.74 | 135.74 | 257.18 | 0.2120 | 31.21 |
DeepI2T | 97.99 | 201.33 | 0.2128 | 31.34 | 136.66 | 260.90 | 0.2124 | 31.23 |
WDR | 97.22 | 190.09 | 0.2162 | 31.98 | 131.57 | 269.00 | 0.2039 | 33.34 |
GCT-TTE | 92.26 | 147.89 | 0.2262 | 30.46 | 107.97 | 169.15 | 0.1961 | 35.17 |