Forecasting Air Quality Pollutants using Ensemble Learning Models

An efficient air quality management system must include ambient air monitoring. It assists administrative bodies in examining the current state and trends of air quality, determining whether the established air quality standards are being violated, identifying non-attainment cities, gaining the knowledge and understanding necessity to develop remedial and preventive measures, and understanding the natural cleansing process taking place in the environment through pollution dilution, dispersion, wind-based movement and other natural occurrences. National authorities employ an Air Quality Index (AQI) number to quantify air pollution levels and inform the public about them. A significant portion of the population will likely experience serious negative health effects as the AQI rises. An air monitor and an air pollutant concentration over a predetermined averaging period are needed for the AQI determination. The results are divided into ranges, and a descriptor is given to each range. A reliable prediction algorithm is required to predict AQI. This study uses data pre-processing, feature selection based on the ANOVA(Analysis Of Variance) f-test, and a variety of ensemble-based classifiers to predict AQI using historical daily ambient air quality data. It offers the best prediction model for AQI detection. The framework's results show promising improvements in AQI prediction.