Carbon Dioxide Emission from Fossil Fuel: A Procedure for Building a Predictive Model

Carbon dioxide is immensely increasing in Bangladesh between the last few decades. The use of fossil fuel is one of the reasons for carbon dioxide emission which is rapidly increasing because of the industrial revolution. The burning of fossil fuels, manufacturing of cement, and the burning of solid, liquid, gas fuels and gas flaring are often responsible for Carbon dioxide emissions. In this paper, we established an emerging model that helps to predict the amount of carbon dioxide emission in upcoming years where the model variables are the reasons to create such emissions. The experiment has been conducted on a data repository named WDI-2015 consisting of the development indicators of Bangladesh produced by the World Bank. At first, we separated the vital indicators from the data source and used multiple stepwise linear regressions to build many regression models. Later, the best model was chosen and evaluated using the acceptable standards of Mean Absolute Error, Relative Absolute Error, and Root Mean Square Error under the capability of k-fold cross-validation. Further, we proved the model significance and performance by comparing it with other classifying methods like Gaussian Process, Decision Table, and Random Forests.