Table 1.
Difference between the predicted value of the model and the actual observed value according to air pollutants.
| Period | Model | CO | NO2 | O3 | PM10 | ||||
|---|---|---|---|---|---|---|---|---|---|
| Conc. (ppm) | Deviation (%) | Conc. (ppm) | Deviation (%) | Conc. (ppm) | Deviation (%) | Conc. (μg/m3) | Deviation (%) | ||
| 2002–2015 | Observed | 0.794 | 0.042 | 0.015 | 54.97 | ||||
| RF | 0.804 | − 1.3 | 0.042 | 0 | 0.014 | 6.7 | 55.34 | − 0.7 | |
| RNN | 0.796 | − 0.3 | 0.041 | 2.4 | 0.015 | 0 | 53.76 | 2.2 | |
| 2016–2020 | Observed | 0.564 | 0.035 | 0.020 | 46.73 | ||||
| RF | 0.714 | − 26.6 | 0.044 | − 25.7 | 0.016 | 20.0 | 41.68 | 10.8 | |
| RNN | 0.576 | − 2.1 | 0.038 | − 8.6 | 0.021 | − 5.0 | 50.84 | − 8.8 | |
RF random forest, RNN recurrent neural network.