Prediction of Air Quality Index and Air Quality Levels in Guwahati-India Using Machine Learning: A Comparative Study

This study provides a deep insight into the factors contributing to air pollution in Guwahati, India. It suggests measures for policymakers and urban planners to develop air quality management plans to control air pollution, not only for the benefit of human health but also for all living beings and the environment. The deteriorating air quality in urban areas, particularly in rapid growing city possesses notable health risks and environmental challenges. The main reason for examining the AQI (Air Quality Index) is the profound effects on health and environmental well-being. This research has analysed the evaluation and prediction of air quality based on the dataset obtained from Kaggle for the period of 2015-2020, which includes data on ten pollutants: PM2.5, PM10, NO, NO2, NOx, NH3, CO, SO2, O3 and Benzene. Three models from ML (Machine Learning), viz. DT (Decision Tree), RF (Random Forest), KNN (K-Nearest Neighbors) have been used for prediction and forecasting the AQI and AQL (Air Quality Index Levels). Finally, it has been observed that the RF Classification showed the highest accuracy in forecasting the AQL and factors such as PM10, PM2.5 and NH3 have been identified as the primary factors in determining AQI rating in Guwahati.