Vehicle CO<sub>2</sub> Emission Prediction Using Deep Learning and Ensemble Machine Learning Methods
Аннотация
Climate change is a global problem, and one of its main causes is global warming, which is predominantly driven by the overwhelming emission of greenhouse gases into the atmosphere, particularly carbon dioxide (CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>). Effective mitigation of global warming requires the monitoring, evaluation, and prediction of greenhouse gas emissions, with particular emphasis on carbon dioxide (CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>). This study is devoted to the application of various machine learning (ML), ensenble and deep learning (DL) models, including MLP, GRU, LSTM, XGBoost, LightGBM, Random Forest, SVR, Decision Tree, and KNN, to predict CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> emissions from vehicles. The models built on a dataset of vehicle emissions with features such as car models, engine sizes, number of cylinders, transmission types, fuel types, and fuel consumptions, are evaluated using MSE, MAE, RMSE, and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> metrics. The results indicate that ensemble learning methods outperform the other evaluated models. Specifically, LightGBM achieved the highest performance with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.92, while the Decision Tree (DT) model demonstrated the lowest performance with an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.85.