Machine learning based analysis for relation between global temperature and concentrations of greenhouse gases
Saloni Kalra, Rishab Lamba, Manoj Sharma
Abstract
Climate change is an important topic that needs to be addressed soon. As the consequences of climate change are extremely serious, such as ocean acidification and extreme weather conditions, it is paramount to learn what are the causes behind the phenomenon to battle it effectively. In this work, authors modeled the relationship between global temperatures and atmospheric concentrations of nitrous oxide, methane, and carbon dioxide on a dataset of 65 years using linear regression, decision tree regression, random forest regression, and Artificial Neural Network. Authors have analyzed the performance of these machine learning algorithms on the data and established that ANN outperforms the other algorithms on the basis of mean square error. Further authors have calculated the importance of each feature (carbon dioxide, methane, and nitrous oxide) using the best performing model-ANN and proved that the contribution of carbon dioxide in the increase of global temperatures is the maximum of the given greenhouse gases.