Comparative Evaluation of Hybrid and Ensemble Machine Learning Models for Air Quality Forecasting in Smart Cities

Jimson A. Olaybar, Lean Joy B. Serot, Jose C. Agoylo Jr., Rolly S. Acaso, Jorton A. Tagud, Alex C. Bacalla

© 2026 Jose C. Agoylo Jr., published by UIKTEN. This work is licensed under the Creative Commons Attribution-NonCommercial 4.0 International. (CC BY-NC 4.0).

Citation Information: SAR Journal. Volume 9, Issue 1, Pages 16-23, ISSN 2619-9955, https://doi.org/10.18421/SAR91-02, March 2026.

Received: 26 January 2026.
Revised: 11 March 2026.
Accepted: 17 March 2026.
Published: 27 March 2026.

Abstract:

Accurate Air Quality Index (AQI) prediction is essential for environmental monitoring and public health management. This study conducts a comparative evaluation of hybrid and ensemble machine learning models using the Beijing Multi-Site Air Quality Dataset, which includes PM₂.₅, PM₁₀, SO₂, NO₂, CO, O₃, and meteorological variables. The CNN–LSTM model achieved the highest overall performance, with an R² of 0.939 and RMSE of 21.04, effectively capturing both temporal and spatial dependencies. The Random Forest model, while slightly less accurate, yielded the lowest MAE (11.95) and offered valuable interpretability through feature importance analysis. Key predictors included PM₂.₅, NO₂, temperature, and humidity. These findings offer a strategic foundation for developing intelligent, real-time air quality monitoring systems in smart cities, balancing predictive precision with computational efficiency and model transparency.

Keywords – Air Quality Index (AQI), CNN–LSTM, machine learning, random forest, smart cities, urban air pollution.

                   

                                                                      Full text PDF