Toward a Machine Learning Approach to Predict the CO2 Rating of Fuel-Consuming Vehicles in Canada

Global warming is becoming a major concern for almost all countries nowadays. Burning fossil fuels, most of which by vehicles is the main cause of increasing toxic CO
2
in the air we breathe. Along with other factors, the emission of CO
2
from vehicles in the environment plays a major role in raising the temperature worldwide. Infrared energy of sunlight by the earth’s surface is absorbed by CO
2
and then re-emitted in all directions causing the greenhouse effect. So, it is very important to take necessary actions to identify high as well as low CO
2
emitting vehicles. The main aim of this work is to predict the CO
2
rating of the vehicle based on various key factors and identify the vehicle with CO
2
emissions beyond the ideal range. For that purpose, the last five years (2017-2021) fuel consumption rating dataset of Canada which is publicly available has been collected and analyzed first. Then, eight machine learning techniques were applied to predict the CO
2
rating based on the key features. From our investigation, we have found that the highest accuracy of 96% is achieved by the random forest algorithm. We hope that this work will contribute to the field of identifying and designing low CO2-emitting vehicles.