Enhancing Urban Air Quality Forecasting: A Comparative Study of Machine Learning Models

Abstract Predicting air quality accurately is essential, especially in fast-growing cities where pollution directly impacts public health and environmental policies. This study compares different machine learning models to forecast air quality in major Indian cities using data from 2015 to 2020. We tested several algorithms, including Extra Trees, Random Forest, Linear Discriminant Analysis, K-Nearest Neighbors, Logistic Regression, and a combination of Random Forest and Extra Trees as an ensemble model. Our results show that the ensemble model consistently provides the most accurate predictions, outperforming individual models by leveraging their strengths. This approach improves the classification of air pollution levels, making it more reliable for real-time air quality monitoring. The findings can help policymakers and environmental experts develop better strategies for reducing pollution and protecting public health. This research highlights the power of machine learning in tackling environmental challenges and improving air quality forecasting.