AQIP: Air Quality Index Prediction Using Supervised ML Classifiers

In current years, Air pollution has emerged as a significant environmental concern. Accuracy modeling the complex relationships between air quality variables using advanced machine learning techniques is a promising area of research. The study aims to evaluate and compare the performance of supervised machine learning methods including Support Vector Regressor (SVR), Random Forest (RF), XGBoost, LightGBM for the prediction of air quality index. For the research, we collect a dataset from Kaggle. To assess the model performance, metrices such as root-mean-square-error (RMSE), Mean Absolute Error (MAE) and coefficient of determination (R2 ) were used. Experimental result showed how LightGBM model outperformed the others in AQI prediction (RMSE = 1.4704, R2 = 0.9987 and MAE = 0.1824). Furthermore, all models were evaluated using these metrices, offering a clear comparison that highlighted the factors contributing to the improved accuracy.