Air Quality Prediction Using Machine Learning: Systematic Literature Review

Air pollution significantly impacts global health and the environment, causing millions of deaths annually. With rapid urbanization and industrialization, especially in countries like China, air quality has deteriorated due to pollutants such as PM2.5, PM10, $\mathrm{CO}_{2}, \mathrm{CO}, \mathrm{NO}_{2}$, and SO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>. Traditional models often fail to predict air quality accurately due to their dependence on optimal parameter sets and data availability. Machine learning (ML) methods, however, offer superior performance by handling large datasets and complex variable relationships. This systematic literature review examines 77 articles from 2020 to 2024, highlighting the most effective ML methods for air quality prediction, including Random Forest Regression (RFR), Gradient Boosting Regression (GBR), Support Vector Machines (SVM), Long Short-Term Memory (LSTM) networks, and LightGBM. The review discusses the types of data used, evaluation metrics, and the models’ performance. The findings emphasize the potential of ML to improve air quality monitoring and prediction, contributing to better public health and environmental management.