Table 8 Results of PM, nox pollutant modeling from stationary and mobile sources using different methods in the grid with a resolution of 1000 m and 500 m (RMSE, MAE in kg/year; R2 is unitless).

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

 

PM (1000 M)

NOx (500 m)

RMSE

MAE

R2

RMSE

MAE

R2

MLP

13.509

5.037

0.90

307.496

50.528

0.90

RF

27.040

8.653

0.56

367.494

62.405

0.86

SVR

43.578

13.994

0.02

1014.953

348.696

−0.13

DT

18.824

3.881

0.81

417.542

44.707

0.81

Back to article page