A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention
Abstract Air pollution is a critical global environmental issue, further exacerbated by rapid industrialization and urbanization. Accurate prediction of air pollutant concentrations is essential for effective pollution prevention and control measures. The complex nature of pollutant data is influenc...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-01-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-88086-1 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832571702252404736 |
|---|---|
| author | Gang Chen Shen Chen Dong Li Cai Chen |
| author_facet | Gang Chen Shen Chen Dong Li Cai Chen |
| author_sort | Gang Chen |
| collection | DOAJ |
| description | Abstract Air pollution is a critical global environmental issue, further exacerbated by rapid industrialization and urbanization. Accurate prediction of air pollutant concentrations is essential for effective pollution prevention and control measures. The complex nature of pollutant data is influenced by fluctuating meteorological conditions, diverse pollution sources, and propagation processes, underscores the crucial importance of the spatial and temporal feature extraction for accurately predicting air pollutant concentrations. To address the challenges of data redundancy and diminished long-term prediction accuracy observed in previous studies, this paper presents an innovative approach to predict air pollutant concentrations leveraging advanced data analysis and deep learning methods. The proposed approach, termed KSC-ConvLSTM, integrates the k-nearest neighbors (KNN) algorithm, spatio-temporal attention (STA) mechanism, the residual block, and convolutional long short-term memory (ConvLSTM) neural network. The KNN algorithm adaptively selects highly correlated neighboring domains, while the residual block, enhanced with the STA mechanism, extracts spatial features from the input data. ConvLSTM further processes the output from STA-ConvNet to capture high-dimensional temporal and spatial features. The effectiveness of the KSC-ConvLSTM approach was validated through predictions of PM2.5 concentrations in Beijing and its surrounding urban agglomeration. The experimental results indicate that the KSC-ConvLSTM approach outperforms benchmark approaches in single-step, multi-step, and trend prediction. It demonstrates superior fitting accuracy and predictive performance. Quantitatively, the proposed KSC-ConvLSTM approach reduces the root mean square error (RMSE) by 4.216–8.458 for prediction averages of 1–12 h of PM2.5 in Beijing, compared with the benchmark approach. The findings show that the KSC-ConvLSTM approach shows considerable potential for predicting, preventing, and controlling air pollution. |
| format | Article |
| id | doaj-art-0e87e6b5cba24250abde4b9a4cb98bab |
| institution | Kabale University |
| issn | 2045-2322 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj-art-0e87e6b5cba24250abde4b9a4cb98bab2025-02-02T12:23:41ZengNature PortfolioScientific Reports2045-23222025-01-0115112010.1038/s41598-025-88086-1A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attentionGang Chen0Shen Chen1Dong Li2Cai Chen3School of Management, Guizhou UniversitySchool of Management, Guizhou UniversitySchool of Architecture and Urban Planning, Beijing University of Civil Engineering and ArchitectureSchool of Architecture and Urban Planning, Beijing University of Civil Engineering and ArchitectureAbstract Air pollution is a critical global environmental issue, further exacerbated by rapid industrialization and urbanization. Accurate prediction of air pollutant concentrations is essential for effective pollution prevention and control measures. The complex nature of pollutant data is influenced by fluctuating meteorological conditions, diverse pollution sources, and propagation processes, underscores the crucial importance of the spatial and temporal feature extraction for accurately predicting air pollutant concentrations. To address the challenges of data redundancy and diminished long-term prediction accuracy observed in previous studies, this paper presents an innovative approach to predict air pollutant concentrations leveraging advanced data analysis and deep learning methods. The proposed approach, termed KSC-ConvLSTM, integrates the k-nearest neighbors (KNN) algorithm, spatio-temporal attention (STA) mechanism, the residual block, and convolutional long short-term memory (ConvLSTM) neural network. The KNN algorithm adaptively selects highly correlated neighboring domains, while the residual block, enhanced with the STA mechanism, extracts spatial features from the input data. ConvLSTM further processes the output from STA-ConvNet to capture high-dimensional temporal and spatial features. The effectiveness of the KSC-ConvLSTM approach was validated through predictions of PM2.5 concentrations in Beijing and its surrounding urban agglomeration. The experimental results indicate that the KSC-ConvLSTM approach outperforms benchmark approaches in single-step, multi-step, and trend prediction. It demonstrates superior fitting accuracy and predictive performance. Quantitatively, the proposed KSC-ConvLSTM approach reduces the root mean square error (RMSE) by 4.216–8.458 for prediction averages of 1–12 h of PM2.5 in Beijing, compared with the benchmark approach. The findings show that the KSC-ConvLSTM approach shows considerable potential for predicting, preventing, and controlling air pollution.https://doi.org/10.1038/s41598-025-88086-1Urban air quality modelingConvolution-sequence hybrid modelSpatial-temporal feature extractionAir pollution predictionDeep learning |
| spellingShingle | Gang Chen Shen Chen Dong Li Cai Chen A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention Scientific Reports Urban air quality modeling Convolution-sequence hybrid model Spatial-temporal feature extraction Air pollution prediction Deep learning |
| title | A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention |
| title_full | A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention |
| title_fullStr | A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention |
| title_full_unstemmed | A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention |
| title_short | A hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio-temporal attention |
| title_sort | hybrid deep learning air pollution prediction approach based on neighborhood selection and spatio temporal attention |
| topic | Urban air quality modeling Convolution-sequence hybrid model Spatial-temporal feature extraction Air pollution prediction Deep learning |
| url | https://doi.org/10.1038/s41598-025-88086-1 |
| work_keys_str_mv | AT gangchen ahybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT shenchen ahybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT dongli ahybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT caichen ahybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT gangchen hybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT shenchen hybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT dongli hybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention AT caichen hybriddeeplearningairpollutionpredictionapproachbasedonneighborhoodselectionandspatiotemporalattention |