Table 5. IAQI and AQI prediction precision in Experiment 4.
From: Air quality historical correlation model based on time series
Predicted item | The optimal input combination | MAPE (%) | RMSE | MAE | Number of hidden states | Matching days |
|---|---|---|---|---|---|---|
IAQIPM10 | IAQIPM10, TSP emissions | 11.35 | 11.45 | 9.36 | 6 | 12 |
IAQIPM2.5 | IAQIPM2.5, TSP emissions | 12.35 | 14.34 | 9.15 | 6 | 3 |
IAQISO2 | IAQISO2,IAQIPM10 & SO2 emissions | 34.89 | 10.03 | 7.59 | 4 | 8 |
IAQINO2 | IAQINO2, IAQISO2, IAQIPM2.5 & NOx emissions | 10.98 | 8.15 | 6.62 | 6 | 15 |
IAQICO | IAQICO, IAQINO2 & NOx emissions | 13.52 | 9.97 | 5.94 | 3 | 12 |
IAQIO3 | IAQIO3, SO2 emissions | 29.87 | 24.12 | 16.26 | 4 | 10 |
AQI | / | 14.47 | 17.59 | 13.59 |
