Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Dec 8;19(24):16494.
doi: 10.3390/ijerph192416494.

Air Quality Modeling with the Use of Regression Neural Networks

Szymon Hoffman  1 , Mariusz Filak  1 , Rafał Jasiński  1
Affiliations

Air Quality Modeling with the Use of Regression Neural Networks

Szymon Hoffman et al. Int J Environ Res Public Health. .

Abstract

Air quality is assessed on the basis of air monitoring data. Monitoring data are often not complete enough to carry out an air quality assessment. To fill the measurement gaps, predictive models can be used, which enable the approximation of missing data. Prediction models use historical data and relationships between measured variables, including air pollutant concentrations and meteorological factors. The known predictive air quality models are not accurate, so it is important to look for models that give a lower approximation error. The use of artificial neural networks reduces the prediction error compared to classical regression methods. In previous studies, a single regression model over the entire concentration range was used to approximate the concentrations of a selected pollutant. In this study, it was assumed that not a single model, but a group of models, could be used for the prediction. In this approach, each model from the group was dedicated to a different sub-range of the concentration of the modeled pollutant. The aim of the analysis was to check whether this approach would improve the quality of modeling. A long-term data set recorded at two air monitoring stations in Poland was used in the examination. Hourly data of basic air pollutants and meteorological parameters were used to create predictive regression models. The prediction errors for the sub-range models were compared with the corresponding errors calculated for one full-range regression model. It was found that the application of sub-range models reduced the modeling error of basic air pollutants.

Keywords: air monitoring; air pollutants; air quality; artificial neural networks; prediction; prediction error; regression.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Figures

Figure 1
Figure 1
Locations of the air monitoring stations Zabrze and Złoty Potok in Poland.
Figure 2
Figure 2
An architecture diagram of the multi-layer perceptron with ten neurons in a single hidden layer.
Figure 3
Figure 3
The scheme of the division into sub-ranges for the Zabrze station.
Figure 4
Figure 4
The scheme of the division into sub-ranges for the Złoty Potok station.
Figure 5
Figure 5
Overall MAE and RMSE values for O3 concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 6
Figure 6
Overall MAE and RMSE values for O3 concentration prediction in PVS models depending on the number of created sub-models, Złoty Potok.
Figure 7
Figure 7
Overall MAE and RMSE values for O3 concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
Figure 8
Figure 8
Overall MAE and RMSE values for O3 concentration prediction in RVS models depending on the number of created sub-models, Złoty Potok.
Figure 9
Figure 9
Overall MAE and RMSE values for NO concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 10
Figure 10
Overall MAE and RMSE values for NO concentration prediction in PVS models depending on the number of created sub-models, Złoty Potok.
Figure 11
Figure 11
Overall MAE and RMSE values for NO concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
Figure 12
Figure 12
Overall MAE and RMSE values for NO concentration prediction in RVS models depending on the number of created sub-models, Złoty Potok.
Figure 13
Figure 13
Overall MAE and RMSE values for NO2 concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 14
Figure 14
Overall MAE and RMSE values for NO2 concentration prediction in PVS models depending on the number of created sub-models, Złoty Potok.
Figure 15
Figure 15
Overall MAE and RMSE values for NO2 concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
Figure 16
Figure 16
Overall MAE and RMSE values for NO2 concentration prediction in RVS models depending on the number of created sub-models, Złoty Potok.
Figure 17
Figure 17
Overall MAE and RMSE values for SO2 concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 18
Figure 18
Overall MAE and RMSE values for SO2 concentration prediction in PVS models depending on the number of created sub-models, Złoty Potok.
Figure 19
Figure 19
Overall MAE and RMSE values for SO2 concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
Figure 20
Figure 20
Overall MAE and RMSE values for SO2 concentration prediction in RVS models depending on the number of created sub-models, Złoty Potok.
Figure 21
Figure 21
Overall MAE and RMSE values for PM10 concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 22
Figure 22
Overall MAE and RMSE values for PM10 concentration prediction in PVS models depending on the number of created sub-models, Złoty Potok.
Figure 23
Figure 23
Overall MAE and RMSE values for PM10 concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
Figure 24
Figure 24
Overall MAE and RMSE values for PM10 concentration prediction in RVS models depending on the number of created sub-models, Złoty Potok.
Figure 25
Figure 25
Overall MAE and RMSE values for CO concentration prediction in PVS models depending on the number of created sub-models, Zabrze.
Figure 26
Figure 26
Overall MAE and RMSE values for CO concentration prediction in RVS models depending on the number of created sub-models, Zabrze.
See this image and copyright information in PMC

References

    1. World Health Organization 7 Million Premature Deaths Annually Linked to Air Pollution. [(accessed on 29 April 2021)]. Available online: www.who.int/mediacentre/news/releases/2014/air-pollution/en.
    1. European Environment Agency . Air Quality in Europe-2020 Report. No. 12/2018. Publications Office of the European Union; Luxembourg: 2020.
    1. Widziewicz Y., Rogula-Kozlowska W., Loska K., Kociszewska K., Majewski G. Health Risk Impacts of Exposure to Airborne Metals and Benzo(a)Pyrene during Episodes of High PM10 Concentrations in Poland. Biomed. Environ. Sci. 2018;31:23–36. doi: 10.3967/bes2018.003. - DOI - PubMed
    1. Trojanowska M., Świetlik R. Heavy metals cadmium, nickel and arsenic environmental inhalation hazard of residents of Polish cities. Environ. Med. 2012;15:33–41.
    1. Gurjar B.R., Molina L.T., Ojha C.S.P. Air Pollution: Health and Environmental Impacts. CRC Press; Boca Raton, FL, USA: 2010. - DOI

Publication types

MeSH terms

LinkOut - more resources