Implementation of machine learning to estimate the air pollutants such as Carbon dioxide, methane, nitrous oxide and total greenhouse gas emissions in Bangladesh

Abstract

Nature and technology are two di erent subject matter with having much dissension between each. Only a few years back, technological growth looked like a threat to nature. However, the bene t of having huge computational power and Machine Learning applications, computers now have the capability of visualizing the vital component of nature. By using the concept of machine learning, researchers have exhibited the limitless use of arti cial intelligence. As a part of that process, we have identi ed a speci c problem on air pollution to tackle by using machine learning that just the human brain is unable to determine. We have taken Bangladesh's harmful emission factors into account, then trained them by using several machine learning techniques like regression and deep learning to predict the emission level. In consequence, we have applied models such as Linear Regression, Long Short Term Memory and Multi- layer Perceptron and found highest 99.05% of accuracy rate also described how this research can be extended in the context of other countries in future years.