Air Quality Index Prediction using Machine Learning

Abstract

Air pollution is a serious worldwide problem which negatively impacts on human health, climate, and ecosystems. Accurate AQI prediction is necessary for timely warning and environmental policy-making. In this paper, we investigate the use of different machine learning algorithms to forecast AQI based on environmental information like pollutant concentrations (PM2.5, PM10, NO₂, CO, SO₂, O₃), temperature, and humidity. The models are trained and tested on real-time and historical air quality datasets. This study compares the algorithms like Linear Regression, RandomForestRegressor, Support Vector Machine (SVM), XGBoostRegressor, GaussianNB. The outcomes reveal that ensemble-based models, specifically Linear Regression and XGBoostRegressor model offer high prediction accuracy.