Detection and Prediction of Air Pollution using Machine Learning and Deep Learning Techniques

Abstract

For the past several years, urbanization and industrialization have been on increase in developed nations, resulting in a huge rise in contaminated air. Citizens and governments have become increasingly worried about the detrimental effects of Air Pollution (AP). Predicting Air Quality (AQ) is a critical step that the government should take because it becoming a major public health concern. The AQ index is a metric for measuring AQ. When fossil fuels such as natural gas, coal, and wood are burned, as well as factories and motor cars, carbon dioxide, nitrogen dioxide, carbon monoxide, and other noxious gases are emitted which leads to AP. This can cause health issues such as cancer, brain impairment, and even death. The forecast of AP allows the government to take preventive actions such as restricting driving hours, partially closing factories, and issuing public service messages. So, the prediction of AP is essential and there is a lot of research being done in this field. This research article reviewed the recent works on AP prediction and present the research using five important sections. In the first section of this review, we will discuss the available methods that can be used to acquire AP data. Second, the processes that can be carried out to pre-process the raw data. Third, the prediction of pollution made use of developing technologies such as Machine Learning (ML) and Deep Learning (DL), and lastly, the evaluation of the model was discussed. The research gap and future steps on the prediction of AP also elaborated. This review will help researchers for a better understanding of the automatic prediction of AP.