Application and Comparison of Machine Learning and Traditional Regression Models for Air Quality Index Prediction in India

Air pollution, a global environmental issue, is a growing concern in developing countries, particularly India. This study analyzes air quality data from 10 major districts of India from 2020-2024, focusing on the impact of seven pollutant indicators on the Air Quality Index (AQI). Data normalization was used to calculate AQI values based on international standards. Three linear regression models were constructed: a full parameter model, one focusing only on particulate matter (PM2.5, PM10), and one excluding other indicators. The experimental results show that the model with particulate matter as a predictor variable outperforms other models, confirming that PM2.5 and PM10 are key indicators for AQI prediction in Indian regions.