Meteorological Factors Governing Particulate Matter Distribution in an Urban Environment

Keywords:
Particulate Matter, Meteorological Factors, Machine Learning, Explainable Artificial Intelligence.

Abstract:
In this study, the impact of meteorological factors on PM10 concentrations in the Belgrade urban area was investigated by using eXtreme Gradient Boosting (XGBoost) and SHapley Additive exPlanations (SHAP) attribution methods. As shown, XGBoost provided reliable PM10 predictions with relative errors in the range from approx. 19% to 26% and correlation coefficients higher than 0.95. The change in emission source intensity, momentum flux intensity, lifted index, humidity, and temperature, as well as concentrations of benzene, NO, NOx and SO2 were the most important variables that described the PM concentration dynamics in Belgrade urban area.

CITATION:

IEEE format

Copy to clipboard

M. Perišić, G. Jovanović, S. Stanišić, A. Šoštarić, A. Stojić, “Meteorological Factors Governing Particulate Matter Distribution in an Urban Environment ,” in Sinteza 2021 - International Scientific Conference on Information Technology and Data Related Research, Belgrade, Singidunum University, Serbia, 2021, pp. 89-93. doi:10.15308/Sinteza-2021-89-93

APA format

Copy to clipboard

Perišić, M., Jovanović, G., Stanišić, S., Šoštarić, A., Stojić, A. (2021). Meteorological Factors Governing Particulate Matter Distribution in an Urban Environment . Paper presented at Sinteza 2021 - International Scientific Conference on Information Technology and Data Related Research. doi:10.15308/Sinteza-2021-89-93

BibTeX format
Download

RefWorks Tagged format
Download