Comparative Analysis of Supervised Machine Learning Techniques for AQI Prediction
Alka Pant, Sanjay Sharma, Mamta Bansal, Mandeep Narang
Abstract
Air pollution is a significant challenge in a populated area. This paper focuses on predicting air quality index using supervised machine learning techniques in the capital city of Uttarakhand state, India, i.e., Dehradun based on the available pollutants (PM<inf>10</inf>, PM<inf>2.5</inf>, SO<inf>2</inf>, NO<inf>2</inf>). The result shows that the decision tree classifier is more accurate, with an accuracy of 98.63%. In contrast, the logistic regression is the least one with an accuracy of 91.78% for air quality prediction. The study also finds that the AQI level is low in May due to high temperatures. The study also finds that the Himalayan drugs-ISBT area is in the poor range of AQI for the capital city of Uttarakhand state.