Artificial Intelligence-Based Framework for Analyzing Crises-Caused Air Pollution

Keywords:
Artificial Intelligence, Machine Learning, Explainable Artificial Intelligence, Metaheuristics, Air Pollution

Abstract:
Understanding the impact of air pollution processes during a crisis is crucial due to the significant risk to human health and to ensure global sustainability. Addressing this issue, this study introduces a novel artificial intelligence-based framework designed to analyze air pollution alterations caused by crises. The framework utilizes seven machine-learning regression models for making predictions: AdaBoost, CatBoost, ExtraTrees, Gradient Boosting, Histogram Gradient Boosting, LightGBM, and XGBoost regressor. Cross-validation is employed to ensure the robustness of the models and to prevent overfitting. The framework includes different metaheuristics algorithms, such as the Firefly Algorithm, Artificial Bee Colony, Harris Hawks Optimization, Sine Cosine Algorithm, Slime Mould Algorithm, and Quantum Superposition Algorithm. The top three performing ensemble models are optimized with the selected metaheuristic algorithm to find the optimal set of hyperparameters and to improve the results. After the optimization process, the best model is selected and evaluated on the dataset, then for explainability, SHAP and SAGE analysis are applied to provide deeper insight into the factors that influence the best model’s predictions. These techniques ensure that the models are not only making precise predictions but also transparent and interpretable, which allows informed decision-making. Finally, the obtained results are visualized interactively for easier analysis of underlying patterns. This study lays the groundwork for a more effective crisis management system to mitigate the adverse of human health and environmental outcomes associated with air pollution caused by crises

CITATION:

IEEE format

Copy to clipboard

T. Bezdan, M. Perišić, G. Jovanović, N. Bačanin Džakula, A. Stojić, “Artificial Intelligence-Based Framework for Analyzing Crises-Caused Air Pollution,” in Sinteza 2024 - International Scientific Conference on Information Technology, Computer Science, and Data Science, Belgrade, Singidunum University, Serbia, 2024, pp. 281-287. doi:10.15308/Sinteza-2024-281-287

APA format

Copy to clipboard

Bezdan, T., Perišić, M., Jovanović, G., Bačanin Džakula, N., Stojić, A. (2024). Artificial Intelligence-Based Framework for Analyzing Crises-Caused Air Pollution. Paper presented at Sinteza 2024 - International Scientific Conference on Information Technology, Computer Science, and Data Science. doi:10.15308/Sinteza-2024-281-287

BibTeX format
Download

RefWorks Tagged format
Download