Air Quality Prediction Using a Machine Learning Hybrid Model in Punjab
Main Article Content
Abstract
Clean air is an essential component for the health and survival of humans and wildlife. Atmospheric pollution has been closely linked to several significant diseases, including cancer, which highlights the critical need to address air quality issues. However, with the rapid pace of industrialization and population growth, human activities such as transportation, household operations, agricultural practices, and industrial processes have significantly contributed to air pollution. This has led to air pollution becoming a major environmental and health concern, particularly in urban areas of developing countries like India. To ensure the maintenance of ambient air quality, it is imperative to conduct regular monitoring and forecasting of air pollution levels. Machine learning has emerged as an innovative and effective technique for predicting the Air Quality Index (AQI) compared to conventional forecasting methods. In this study, we applied AQI prediction methods to Punjab, India. The research focused on analysing 11 air contaminants and 9 meteorological parameters over a comprehensive timeframe spanning from July 2019 to September 2023. To achieve accurate predictions, several advanced machine learning models were employed, including LightGBM, Random Forest, Catboost, Adaboost, and XGBoost. Among these models, the Catboost model demonstrated superior performance, achieving an exceptionally high R² correlation coefficient of 0.9997. It also recorded a mean absolute error (MAE) of 0.62, a mean square error (MSE) of 0.60, and a root mean square error (RMSE) of 0.78, making it the most effective predictor in this study. On the other hand, the Adaboost model exhibited the least predictive capability, with an R² correlation coefficient of 0.9471.
Article Details

This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.