Machine Learning for Air Pollution

Machine Learning has revolutionized the methods employed for various prediction problems. Air pollution has been a significant problem for a long time. One of the most crucial natural resources is air and infiltration of air has resulted in various threats to our environment and is concerned with the survival of various living beings. This paper aims to throw some light on various research and development in air pollution epidemiology incorporating the advantages of machine learning over various conventional methods which usually employ several mathematical and statistical approaches which generally prove to be more complex and less accurate. A review of various standards for evaluating air quality for a given area and some machine learning models have been presented here. There are six criteria pollutants listed by NAQQS that are considered when evaluating air quality for a region. The machine learning models discussed are RF (Random Forest), SVM (Support vector machine), GA (Genetic Algorithm), Decision Trees, LS-SVM (Least Squares Support Vector Machine), and ANN (Artificial Neural Network). Along with this learning paradigms (Prediction-based and knowledge discovery methods) required have also been discussed here. In addition to this, some issues and prospects related to this matter have also been highlighted here.