№79-24

Methods for forecasting air pollution based on machine learning

P. Lomazov¹, А. Pavlуchenko¹, Yu. Buchavyi¹

¹Dnipro University of Technology, Dnipro, Ukraine

Coll.res.pap.nat.min.univ. 2024, 79:278–291

Full text (PDF)

https://doi.org/10.33271/crpnmu/79.278

ABSTRACT

Goal. To investigate modern methods for predicting air pollution using machine learning algorithms.

Methodology. A modeling approach was used to develop forecasting algorithms that consider the temporal and spatial characteristics of the data. The method of specification was applied to refine the dependencies between variables. Statistical generalization methods were also employed to remove noise, fill in missing values, and identify trends and anomalies. Additionally, this study examines the Random Forest, LSTM, GBM, SVM, and MLP algorithms, discussing their advantages, limitations, and potential applications in environmental monitoring.

Research results. A correlation analysis of pollutant concentrations and meteorological parameters was conducted using data from an observation station in Dnipro. The feasibility of applying machine learning methods for analyzing and forecasting time series of pollutant concentrations was substantiated. The results confirm that machine learning algorithms enable high forecasting accuracy by processing large datasets and capturing complex relationships between pollution sources and meteorological conditions. The Random Forest and GBM algorithms proved effective for data with static dependencies, whereas LSTM was optimal for time series modeling.

Scientific novelty. The patterns of the influence of atmospheric parameters on the accuracy of pollution prediction by machine learning algorithms have been established. It has been found that models such as Random Forest, LSTM, GBM, etc., exhibit different sensitivity to the volume and quality of input data. These dependencies contribute to improving forecasting methodology, ensuring higher accuracy and adaptation to changing conditions.

Practical significance. The developed models can be applied in urban management and environmental monitoring. Specifically, they can be integrated into early warning systems for informing citizens about pollution risks, optimizing traffic flows, and supporting mobile applications and web platforms that provide real-time air quality information.

Keywords: machine learning, Random Forest, LSTM, meteorological factors, forecasting, IoT technologies, monitoring.

References

1. Mishchuk, O.S. (2019). Pidvyshchennia tochnosti prohnozuvannia parametriv zabrudnennia povitria. U Komp’iuterne modeliuvannia ta optymizatsiia skladnykh system (KMOSS2019): materialy V-i mizhnarodnoi naukovo-tekhnichnoi konferentsii, 129–130.

2. Vyshnevskyi, A.V., & Potapova, N.A. (2024). Metody analizu danykh v informatsiinii systemi prohnozuvannia pokaznykiv zabrudnennia povitria. Prykladni aspekty suchasnykh mizhdystsyplinarnykh doslidzhen, 156–157.

3. Ameer, S., Shah, M. A., Khan, A., Song, H., Maple, C., Islam, S. U., & Asghar, M. N. (2019). Comparative Analysis of Machine Learning Techniques for Predicting Air Quality in Smart Cities. IEEE Access, 7, 128325–128338. https://doi.org/10.1109/access.2019.2925082

4. Zaporozhets, A. O. (2017). Analiz zasobiv monitorynhu zabrudnennia povitria navkolyshnoho seredovyshcha. Science-Based Technologies, 35(3), 242–252.

5. Renugadevi, R., Vyshnavi, T., Prasanna, T. R., & Lahari, P. S. (2023). Air quality prediction using random forest algorithm. International Conference on Research Methodologies in Knowledge Management, Artificial Intelligence and Telecommunication Engineering (RMKMATE), 1–6. https://doi.org/10.1109/RMKMATE59243.2023.10369180

6. Xayasouk, T., Lee, H., & Lee, G. (2020). Air pollution prediction using long short-term memory (LSTM) and deep autoencoder (DAE) models. Sustainability, 12(6), 2570. https://doi.org/10.3390/su12062570

7. Su, Y. (2020). Prediction of air quality based on gradient boosting machine method. International Conference on Big Data and Informatization Education (ICBDIE), Zhangjiajie, China, 395–397. https://doi.org/10.1109/ICBDIE50010.2020.00099

8. Widyarini, L., & Purnomo, H. (2024). Air quality prediction using the support vector machine algorithm. Journal of Information Systems and Informatics, 6(2), 652–661. https://doi.org/10.51519/journalisi.v6i2.705

9. Surya, R., Jeevananthan, P., Muniyappan, A., Prakash, P. K. B., & Saveeth, P. (2024). Air Quality Monitoring and Prediction for Industries. 2024 International Conference on Science Technology Engineering and Management (ICSTEM), 1–6. https://doi.org/10.1109/icstem61137.2024.10560996

10. Widyarini, L., & Purnomo, H. (2024). Air quality prediction using the support vector machine algorithm. Journal of Information Systems and Informatics, 6(2), 652–661. https://doi.org/10.51519/journalisi.v6i2.705

11. Osadchuk, O., Osadchuk, Ya., & Dovhun, V. (2024). Dosiahnennia ta vyklyky khmarnykh tekhnolohii monitorynhu yakosti povitria. Measuring and Computing Devices in Technological Processes, (2), 18–25. https://doi.org/10.31891/2219-9365-2024-78-2

12. Freeman, B. S., Taylor, G., Gharabaghi, B. J., & Thé, J. (2018). Forecasting air quality time series using deep learning. Journal of the Air & Waste Management Association, 68(8), 866–886. https://doi.org/10.1080/10962247.2018.1459956