Related Experiment Video
Updated: Jan 12, 2026
04:04
Asthma Detection Research Based on Voice Signal Processing and Machine Learning
Published on: July 22, 2025
902
RETRACTED: Interpretable machine learning framework for predicting Urban air quality
Rana Muhammad Amir Latif1, Tahir Iqbal2, Ismaeel Abdel Qader3
1The Center for Modern Chinese City Studies, School of Geographic Sciences, East China Normal University, Shanghai, China.
Plos One
|November 7, 2025
Summary
Machine learning models effectively forecast air quality. Random Forest and XGBoost showed the best performance in predicting the Air Quality Index (AQI), offering insights for urban air pollution management.
Area of Science:
- Environmental Science
- Computer Science
- Public Health
Background:
- Urban air pollution poses significant risks to public health and environmental sustainability.
- The Air Quality dataset from the UCI ML Repository, though dated (2004-2005), remains a valuable benchmark for evaluating air quality forecasting methods.
- Machine learning (ML) offers potential solutions for predicting air quality and informing policy.
Purpose of the Study:
- To evaluate the predictive performance of five machine learning models (LR, DT, RF, XGBoost, SVR) for Air Quality Index (AQI) forecasting.
- To identify the most influential features for AQI prediction.
- To develop an interpretable and reproducible ML framework for air quality management.
Main Methods:
- Utilized the UCI ML Repository's Air Quality dataset, performing pre-processing, feature engineering, and chronological splitting.
- Trained and rigorously tuned five ML models: Linear Regression (LR), Decision Tree (DT), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Support Vector Regression (SVR).
- Assessed model performance using RMSE, MAE, and R2, with statistical significance confirmed via bootstrap confidence intervals and t-tests. Employed SHAP for interpretability.
Main Results:
- Ensemble models, specifically Random Forest and XGBoost, demonstrated superior performance in AQI forecasting.
- Random Forest achieved the lowest Root Mean Squared Error (RMSE) of 12.48 and Mean Absolute Error (MAE) of 9.35.
- XGBoost yielded the highest coefficient of determination (R2) of 0.89. NOx, PM2.5, and CO were identified as key predictors.
Conclusions:
- Interpretable machine learning models provide a reproducible and efficient framework for AQI forecasting.
- The study underscores the value of benchmark datasets for validating ML methodologies in environmental science.
- Findings support the application of ML for smart city air quality management and public health policy development.
