Predicting Air Quality Using Machine Learning: A Comparative Analysis

Air pollution is a significant global issue, posing severe risks to public health, ecosystems, and climate stability. It is primarily caused by industrial emissions, vehicle exhaust, and the burning of fossil fuels, leading to elevated levels of harmful pollutants like particulate matter PM<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">10</inf>, nitrogen dioxide NO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>, and sulfur dioxide SO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>. The adverse effects of air pollution include respiratory diseases, cardiovascular problems, and environmental degradation, necessitating urgent action to monitor, predict, and mitigate its impact. Machine learning (ML) plays a crucial role in predicting air pollution by analyzing large, complex datasets to identify patterns and trends in pollutant levels. It enables accurate and timely predictions, supporting decision-makers in implementing proactive measures to mitigate air quality issues and protect public health. The novelty of this paper lies in its comprehensive evaluation of three widely used machine learning models—Random Forest (RF), Support Vector Classifier (SVC), and K-Nearest Neighbors (KNN)—for predicting the Air Quality Index (AQI) using multiple performance metrics, including accuracy, recall, precision, and F-measure. Among the models, RF demonstrated near-perfect results, achieving 99.99% across all metrics. This highlights the robustness and reliability of the RF model in handling complex datasets with high-dimensional features and intricate patterns, demonstrating its potential as an effective solution for air quality prediction when compared to traditional techniques.