AIR QUALITY INDEX PREDICTION USING PYTHON BASED NEURAL NETWORKS

Abstract

Air pollution has become a major global concern, significantly impacting public health and environmental sustainability. Accurate
prediction of the Air Quality Index (AQI) can assist authorities in implementing timely preventive actions and improving air
management systems. This study presents a Python-based predictive framework employing artificial neural networks (ANNs) to
forecast AQI levels using real-time environmental data. The proposed model integrates multiple pollutant parameters, including
PM2.5, PM10, NO₂, SO₂, CO, and O₃, along with meteorological factors such as temperature and humidity. Data preprocessing
techniques—such as normalization, feature scaling, and outlier removal—are applied to enhance model reliability. The neural
network is designed and trained using TensorFlow and Keras libraries to capture complex nonlinear relationships among
pollutants. Performance evaluation using metrics like Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R²
demonstrates that the model achieves high accuracy in predicting AQI values across different regions. The results indicate that the
proposed neural network–based system can serve as an efficient tool for early air quality forecasting and environmental decisionmaking, offering a foundation for future smart city and IoT-based air monitoring applications.