Table 2 Performance of different models on testing and training Sets.

From: PM2.5 concentration 7-day prediction in the Beijing–Tianjin–Hebei region using a novel stacking framework

Model

Training set

Testing set

R2

(µg/m³)

MAE

(µg/m³)

RMSE

(µg/m³)

MAPE

(%)

R2

(µg/m³)

MAE

(µg/m³)

RMSE

(µg/m³)

MAPE

(%)

LSTM

0.98

2.64

3.77

0.13

0.94

7.85

9.23

0.31

RF

0.99

1.17

2.00

0.05

0.93

7.66

9.34

0.25

KNN

0.95

3.74

5.80

0.13

0.76

12.42

17.84

0.34

ML

0.96

3.74

5.44

0.24

0.93

7.95

10.03

0.47

Stacking

0.97

3.26

4.69

0.19

0.96

6.08

7.74

0.26

Back to article page