An Ensemble based Machine Learning Approach for Air Pollution Detection and Control: A Case Study

The rapid advancement of technology and the expansion of industrial activities have significantly compromised air quality, adversely affecting the overall quality of life. Air pollution shows a major environmental challenge with increased consequences negatively affecting human health, ecological balance and economic progress. This leads to a necessity for precise detection and monitoring of air pollution levels. This work uses a dataset of AQI (Air Quality Index) of Hyderabad, the capital city of Telangana, and analyses the level of air pollution in the city. Air Pollution Detection helps in finding the level of air pollution in Hyderabad, where the dataset is split into test and train sets and Machine Learning algorithms are performed. In contrast to traditional estimation algorithms, including statistical and numerical methods, Machine Learning techniques offer enhanced accuracy and efficiency in air quality monitoring systems. In this study, several Machine Learning models, namely Ensemble XGBoost-Random Forest, Gradient Boosting, Random Forest, and XGBoost, were considered for predicting the levels of air pollution. These models use computational power afforded by ML algorithms that could analyse complex datasets and discern patterns and trends that can slip through the sieve of more traditional statistical methods. Using the data covering all environmental factors, examples include particulate matter, ozone, nitrogen dioxide, meteorological variables, and other relevant factors, predictive models are provided to detect the level of air pollution. With careful comparison and evaluation, Ensemble Gradient Boosting comes out as the best model that predicts well the level of air pollution with higher accuracy and reliability. The standalone gradient boosting algorithm gives an MSE of 0.0000998544 and R2 of 0.99737955298. Major Findings: This paper evaluates ensemble ML models (XGBoost-RF, Gradient Boosting, etc.) for Hyderabad’s AQI prediction, with Gradient Boost

View article page

Download PDF