University of Oulu

Hulkkonen, M., Lipponen, A., Mielonen, T., Kokkola, H., & Prisle, N. L. (2022). Changes in urban air pollution after a shift in anthropogenic activity analysed with ensemble learning, competitive learning and unsupervised clustering. Atmospheric Pollution Research, 13(5), 101393.

Changes in urban air pollution after a shift in anthropogenic activity analysed with ensemble learning, competitive learning and unsupervised clustering

Saved in:
Author: Hulkkonen, Mira1; Lipponen, Antti2; Mielonen, Tero2;
Organizations: 1Nano and Molecular Systems Research Unit, P.O. BOX 8000, University of Oulu, FI-90014, Oulu, Finland
2Finnish Meteorological Institute, Atmospheric Research Centre of Eastern Finland, FI-70211, Kuopio, Finland
3Center for Atmospheric Research, P.O. BOX 4500, University of Oulu, FI-90014, Oulu, Finland
Format: article
Version: published version
Access: open
Online Access: PDF Full Text (PDF, 3.6 MB)
Persistent link:
Language: English
Published: Elsevier, 2022
Publish Date: 2022-09-06


Urban air pollution is a health hazard linked to anthropogenic emissions. Reliable evaluation of changes in pollutants due to altered emissions requires considering meteorological and other variability influencing concentrations. Here, a combination of ensemble learning, competitive learning and unsupervised clustering is proposed and applied to leverage the change analysis of particulate matter (PM2.5) and other pollutants.

Machine Learning (ML) algorithms Random Forest (RF) and Self-Organizing Map (SOM) were trained with historical meteorological data, pollutant concentrations and traffic indicators. The importance of different variables for local PM2.5 was determined with RF. SOM was configured for multivariable cluster analysis. The trained SOM enabled predicting a cluster for new data representing conditions with shifted anthropogenic activity. The prediction forms a benchmark for the analysed period with maximized meteorological similarity, which facilitates identifying changes in ambient pollutants due to changed emissions.

The method was applied to data from the start of COVID-19 pandemic, 3/2020, when emissions suddenly decreased. For measurements from Helsinki, Finland, the SOM yielded a statistically significant change in PM2.5 (−0.7%), NO2 (−33%) and O3 (+17%). Comparing data from 3/2020 to data from 3/2017–2019 produced different results (PM2.5 −1.7%, NO2 −37%, O3 −4.0%). Statistical indicators confirmed better compatibility between the analysed period and its benchmark when using the SOM prediction instead of calendar-based selection: Average RMSRE was 19%-points lower and Willmott’s dr 41% higher with SOM than with 3/2017–2019.

Based on the case study and method evaluation, using ML for multivariate analysis of changed air pollution is feasible and yields meaningful results.

see all

Series: Atmospheric pollution research
ISSN: 1309-1042
ISSN-E: 1309-1042
ISSN-L: 1309-1042
Volume: 13
Issue: 5
Article number: 101393
DOI: 10.1016/j.apr.2022.101393
Type of Publication: A1 Journal article – refereed
Field of Science: 1172 Environmental sciences
Funding: This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program, project SURFACE (grant agreement no. 717022). The authors also gratefully acknowledge the financial contribution from the Academy of Finland, including grant Nos. 308292, 316743, 308238, 314175, and 335649.
EU Grant Number: (717022) SURFACE - The unexplored world of aerosol surfaces and their impacts.
Academy of Finland Grant Number: 316743
Detailed Information: 316743 (Academy of Finland Funding decision)
308238 (Academy of Finland Funding decision)
314175 (Academy of Finland Funding decision)
335649 (Academy of Finland Funding decision)
Dataset Reference: Supplementary data to this article can be found online at
Copyright information: © 2022 Turkish National Committee for Air Pollution Research and Control. This is an open access article under the CC BY license (