Deep Learning-Driven Predictive Models for Real-Time Urban Air Quality Assessment Using Big Data Frameworks
DOI:
https://doi.org/10.5281/zenodo.20428558Keywords:
Artificial Intelligence; Predictive Analytics; Air Quality Monitoring; Big Data; Urban Health; Machine Learning; Decision-Making SupportAbstract
Abstract
Environmental factors are responsible for 28% of premature deaths in the European Union, notably due to urban air pollution. INQUA, the global network of terrestrial researchers investigating the impacts of air quality, greenhouse gas emissions, and climate change on human health, responds to these threats with an open-ended challenge to develop reliable predictive models for urban air quality monitoring. Such externalities are increasingly treated as a common good. The approach promises to improve the public health decision-making process and governance by enabling predictive analytics and beyond-the-air-data-informed decision-support systems. Empirical evidence is provided by two pilot studies, one in a megacity with high-density urban cores and the other in a mid-sized city.
Resistance to predictive urban air quality models stems from concerns about the accuracy of traditional regression estimates when compared with modern supervised learning. The ensemble of data from real-time sensor networks, remotely sensed earth properties, and urban activity emissions should determine both algorithms for accurate future estimates and feature-engineering procedures to identify the most effective and interpretable paths for each pollutant. Well-designed learning tasks improve predictive skills, providing a probabilistic assessment that accounts for quality uncertainty, reinforced by temporal validation tests and quantile-class predictive-distribution measures.
References
[1]Govea, J., Gaibor-Naranjo, W., Sanchez-Viteri, S., & Villegas-Ch, W. (2024). Integration of data and predictive models for the evaluation of air quality and noise in urban environments. Sensors, 24(2), 311.
[2]Zayed, R., & Hassan, M. (2024). Air quality index prediction using DNN-Markov modeling for healthcare applications. Applied Artificial Intelligence, 38(1), 1–18.
[3]Omoniyi, S., & Ajayi, O. (2024). Exploratory data analytics of air pollutant data for air quality management application. International Journal of Technology and Systems, 9(2), 82–107.
[4]Sembiring, I., Manongga, D., Rahardja, U., & Aini, Q. (2024). Understanding data-driven analytic decision making on air quality monitoring: An empirical study. Advanced Trends in Technology, 6(3), 418–431.
[5]Hettige, K. H., Ji, J., Xiang, S., Long, C., Cong, G., & Wang, J. (2024). AirPhyNet: Harnessing physics-guided neural networks for air quality prediction. arXiv preprint.
[6]Panja, M., Chakraborty, T., Biswas, A., & Deb, S. (2024). Extreme spatiotemporal graph convolutional networks for air quality forecasting. arXiv preprint.
[7]Khan, H., Tso, J., Nguyen, N., Kaushal, N., Malhotra, A., & Rehman, N. (2024). Attention-enhanced deep multitask spatiotemporal learning for air quality prediction. arXiv preprint.
[8]Semlali, B. E. B., El Amrani, C., Ortiz, G., Boubeta-Puig, J., & Garcia-de-Prado, A. (2024). SAT-CEP-monitor: Air quality monitoring architecture combining complex event processing and satellite sensing. arXiv preprint.[9]
[9]Rosca, C. M., Popescu, D., & Ionescu, B. Data-driven approaches for predicting and forecasting urban air quality index. Applied Sciences, 15(8), 4390.
[10]Varshney, S., Shrivastava, J. N., & Gupta, N. Transforming air quality index prediction using machine learning: Insights from the Taj Trapezium Zone. Engineering, Technology & Applied Science Research, 15(5), 26741–26749.
[11]Latif, R. M. A., et al. Interpretable machine learning framework for urban air quality prediction. Environmental Technology & Innovation, 31, 103000.
[12]Berkani, S., Gryech, I., Ghogho, M., Guermah, B., & Kobbane, A. (2023). Data-driven forecasting models for urban air pollution: MoreAir case study. IEEE Access, 11, 133131–133142.
[13]Udristioiu, M. T., El Mghouchi, Y., & Yildizhan, H. (2023). Hybrid machine learning models for PM and AQI prediction. Journal of Cleaner Production, 421, 138496.
[14]Neo, E. X., Hasikin, K., Lai, K. W., Mokhtar, M. I., & Azizan, M. M. (2023). Artificial intelligence-assisted air quality monitoring for smart city management. PeerJ Computer Science, 9, e1306.
[15]Sekar, S., et al. Hybrid convolutional attention LSTM for PM2.5 prediction using big data analytics. Scientific Reports.
[16]Ansari, A., & Singh, P. (2024). Machine learning trends in air pollution prediction: A bibliometric analysis. Environmental Research and Technology.
[17]Espinosa, R., et al. (2024). Multi-criteria time series forecasting for air quality prediction. Applied Soft Computing, 113, 107850.
[18]Zhao, G., Huang, G., Hong, H., He, H., & Ren, J. (2024). Regional spatiotemporal collaborative models for air quality prediction. IEEE Access, 12, 134903–134919.
[19]Xu, X., & Yoneda, M. (2024). Multitask air-quality prediction using LSTM-autoencoder models. IEEE Transactions on Cybernetics, 54(5), 2577–2586.
[20]Rambha, U. B., & Seshashayee, M. (2024). Time series augmentation using VAR and LSTM for air quality prediction. International Journal of Mechanical Engineering, 9(4), 1–11.
Additional Files
Published
Data Availability Statement
The document references data from sensor networks in Nashik, Pune, Santiago de Chile, and Los Angeles, processed using the M7 data-processing framework, but no formal data availability statement or repository link is included.
