Table 4 Modeling results of various methods for PM10, PM2.5, and nox with considering auto-correlation of the pollutants (RMSE, MAE in µg/m3; R2 is unitless).

From: Unravelling the importance of spatial and temporal resolutions in modeling urban air pollution using a machine learning approach

 

PM10 (3 days)

PM2.5 (6 days)

NOx (4 days)

RMSE

MAE

R2

RMSE

MAE

R2

RMSE

MAE

R2

MLP

68.590

35.569

0.47

37.299

18.479

0.47

12.104

7.747

0.36

RF

72.389

34.333

0.41

38.733

19.165

0.42

13.346

8.016

0.22

Back to article page