Machine Learning-Based Prediction of Carbon Dioxide Emissions from Automobiles and Influencing Factors

Abstract

In recent years, the number of automobiles has been increasing globally, further leading to an increase in carbon dioxide emissions year after year. Carbon dioxide emissions from automobiles have become an important factor in global climate change and have attracted global attention. Firstly, the data were analyzed for missing values all duplicates were removed, and the main discrete random variables were converted into continuous random variables; then, the correlation coefficients of Pearson correlation coefficients were used to analyze the correlation between each automobile characteristic, and heat maps were drawn to remove the influencing factors with weak correlation; finally, the prediction models were established based on multiple linear regression and the Random Forest method, respectively. The results show that in both models, Fuel Consumption Comb is the most important factor influencing the growth of automobile carbon dioxide emissions; the random forest model is better than the multiple linear regression model and can effectively predict automobile carbon dioxide emissions.