Prediction of PM 2.5 concentration based on a CNN-LSTM neural network algorithm
PeerJ · 06.08.2024 · 12 · e17811
DOI: 10.7717/peerj.17811 ↗Kurz gefasst
Fachartikel in der Zeitschrift PeerJ, online erschienen am 06.08.2024. Verfasst von Xuesong Bai und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Information and Control Engineering, Qingdao University of Technology, Qingdao, Qingdao City, Shandong Province, China. Hauptthemen laut MEDLINE: Particulate Matter, Neural Networks, Computer, Air Pollutants. Die Arbeit wurde bisher 40-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Youth Foundation of Shandong Natural Science, Foundation of Chinese Academy of Sciences, Key Research and Development Program for Shandong.
Abstract (Originalsprache)
Fine particulate matter (PM 2.5 ) is a major air pollutant affecting human survival, development and health. By predicting the spatial distribution concentration of PM 2.5 , pollutant sources can be better traced, allowing measures to protect human health to be implemented. Thus, the purpose of this study is to predict and analyze the PM 2.5 concentration of stations based on the integrated deep learning of a convolutional neural network long short-term memory (CNN-LSTM) model. To solve the complexity and nonlinear characteristics of PM 2.5 time series data problems, we adopted the CNN-LSTM deep learning model. We collected the PM 2.5 data of Qingdao in 2020 as well as meteorological factors such as temperature, wind speed and air pressure for pre-processing and characteristic analysis. Then, the CNN-LSTM deep learning model was integrated to capture the temporal and spatial features and trends in the data. The CNN layer was used to extract spatial features, while the LSTM layer was used to learn time dependencies. Through comparative experiments and model evaluation, we found that the CNN-LSTM model can achieve excellent PM 2.5 prediction performance. The results show that the coefficient of determination (R 2 ) is 0.91, and the root mean square error (RMSE) is 8.216 µg/m 3 . The CNN-LSTM model achieves better prediction accuracy and generalizability compared with those of the CNN and LSTM models (R 2 values of 0.85 and 0.83, respectively, and RMSE values of 11.356 and 14.367, respectively). Finally, we analyzed and explained the predicted results. We also found that some meteorological factors (such as air temperature, pressure, and wind speed) have significant effects on the PM 2.5 concentration at ground stations in Qingdao. In summary, by using deep learning methods, we obtained better prediction performance and revealed the association between PM 2.5 concentration and meteorological factors. These findings are of great significance for improving the quality of the atmospheric environment and protecting public health.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- School of Information and Control Engineering, Qingdao University of Technology, Qingdao, Qingdao City, Shandong Province, China.
- Key Laboratory for Semi-Arid Climate Change of the Ministry of Education, College of Atmospheric Sciences, Lanzhou University, Lanzhou, Lanzhou City, Gansu Province, China.
Förderung
- Youth Foundation of Shandong Natural Science
- Foundation of Chinese Academy of Sciences
- Key Research and Development Program for Shandong
- The Youth Foundation of Shandong Natural Science
- The Foundation of Chinese Academy of Sciences
Zitieren
APA
Bai, X., Zhang, N., Cao, X., & Chen, W. (2024). Prediction of PM 2.5 concentration based on a CNN-LSTM neural network algorithm. PeerJ, 12, e17811. https://doi.org/10.7717/peerj.17811
Vancouver
Bai X, Zhang N, Cao X, Chen W. Prediction of PM 2.5 concentration based on a CNN-LSTM neural network algorithm. PeerJ. 2024;12:e17811. doi:10.7717/peerj.17811
BibTeX
@article{Bai2024,
author = {Bai, Xuesong and Zhang, Na and Cao, Xiaoyi and Chen, Wenqian},
title = {Prediction of PM 2.5 concentration based on a CNN-LSTM neural network algorithm},
journal = {PeerJ},
year = {2024},
volume = {12},
pages = {e17811},
doi = {10.7717/peerj.17811}
}
EUVIMED