Predicting Air Pollution Levels in Pune, India using Generative Adversarial Networks

Abstract

Fuel combustion, industrial and factory exhausts, and mining activities contribute to air pollution. Predicting and evaluating the quality of air is a field of study that is growing in importance. This research builds a Generative Adversarial Network (GAN) air quality prediction model. A pre-trained accurate model was applied to predict pollutant levels in air at a given location based on historical data. The prediction GAN model utilized pollutants datasets of Particulate matter (PM2.5 and PM10), Nitrogen dioxide (NO2), Carbon monoxide (CO), and Ozone (O3) between 2016 and 2021 in Pune, India. The Root Mean Square Error (RMSE) statistical measure was used to assess the model's performance accuracy. The close alignment between real and predicted values underscores the high precision of the GAN model in forecasting air pollutant levels.