Predictive modeling of air pollution levels: A state-of-the-art review of machine learning techniques

Abstract

Air pollution remains a significant environmental concern, necessitating accurate monitoring and forecasting techniques. This survey paper reviews state-of-the-art machine learning approaches for air quality prediction, including artificial intelligence, decision trees, deep learning, and ensemble methods. It examines data sources, preprocessing techniques, and the core algorithms employed across different pollutants and regions. The primary objective of this survey paper is to conduct a comprehensive examination of various big data analytics and machine learning methodologies that have been employed for the purpose of forecasting air quality levels. The paper provides an in-depth review and synthesis of existing published research studies that have utilized artificial intelligence techniques, decision tree algorithms, deep learning models, and other advanced approaches to evaluate and predict air quality indicators. Additionally, the survey sheds light on the current challenges faced in this domain and identifies potential areas that necessitate further investigation and research efforts.