Artificial Intelligence Approaches for Predicting Air Quality Index and Air Quality Grades: A Comparative Study

Abstract

Air pollution has become a critical environmental and public-health concern, creating a need for accurate and timely assessment of air-quality conditions. The Air Quality Index (AQI) provides a standardized numerical representation of atmospheric pollution and converts complex pollutant information into understandable air-quality grades. This review examines Artificial Intelligence (AI) approaches for predicting AQI values and corresponding air-quality grades using diverse environmental datasets. It discusses major pollutants, including PM2.5, PM10, NO2, SO2, CO, O3, and NH3, along with meteorological parameters such as temperature, relative humidity, wind speed, and wind direction. The review further explores machine learning and deep learning techniques for identifying nonlinear relationships, temporal variations, pollutant patterns, and changing air-quality conditions. Data sources and preprocessing methods, including monitoring stations, low-cost sensors, mobile monitoring, satellite observations, calibration, missing-value treatment, and anomaly detection, are also considered. In addition, the study highlights integrated frameworks combining multi-sensor networks, data fusion, atmospheric information, IoT-enabled monitoring, and AI-based assessment. Challenges related to data quality, nonlinear interactions, spatial-temporal variability, computational complexity, and model reliability are discussed. Overall, AI-driven AQI prediction can effectively support real-time monitoring, health-risk communication, environmental management, pollution-control planning, and timely public-health decision-making