Table 1 RMSE of predicted results with different hyperparameters.
From: An adaptive adjacency matrix-based graph convolutional recurrent network for air quality prediction
Pollutant | Correlation threshold | Time lag | |||||
|---|---|---|---|---|---|---|---|
24 | 48 | 72 | 96 | 120 | 144 | ||
PM2.5 | 0.7 | 0.00711 | 0.00710 | 0.00743 | 0.00722 | 0.05002 | 0.01405 |
0.75 | 0.04749 | 0.04646 | 0.00700 | 0.04844 | 0.00706 | 0.00948 | |
0.8 | 0.06249 | 0.00954 | 0.00932 | 0.00975 | 0.00888 | 0.01369 | |
PM10 | 0.7 | 0.00871 | 0.00897 | 0.00855 | 0.00884 | 0.01710 | 0.01790 |
0.75 | 0.00884 | 0.00881 | 0.00922 | 0.00900 | 0.01361 | 0.01402 | |
0.8 | 0.01741 | 0.01084 | 0.01104 | 0.01195 | 0.01460 | 0.01862 | |
O3 | 0.7 | 0.01687 | 0.02389 | 0.07028 | 0.06101 | 0.07809 | 0.01690 |
0.75 | 0.05828 | 0.05542 | 0.10183 | 0.06369 | 0.01572 | 0.02537 | |
0.8 | 0.12418 | 0.01838 | 0.02919 | 0.06860 | 0.02184 | 0.02682 | |
