Carbon Footprint: Machine Learning Models for Prediction of Missing Values

The lack of information on the reason for the evolution of greenhouse gas emissions does not help policy makers to develop the suitable roadmaps to contribute in the reduction of these emissions. In this paper, our main objective is to prove how we can use some Machine Learning models to predict effectively unreported corporate greenhouse gas (GHG) emissions. The used model is the Linear Regression. The experimentations are conducted on the 2020 CDP dataset, using several features: country, city, organization, population and source of carbon emissions. The results show that these emissions can be predicted with high accuracy.