Firman Ridwan Petervan Siburian, Suharjito

Boosting-Based Machine Learning Models and Hyperparameter Tuning for Predicting Vehicle Carbon Dioxide Emission

Introduction

Boosting-based machine learning models and hyperparameter tuning for predicting vehicle carbon dioxide emission. Predict vehicle CO2 emissions using boosting-based machine learning models. Compares GBM, CatBoost, XGBoost, with XGBoost achieving superior performance. Supports sustainable transport policy.

Abstract

Sustainable development and climate change are central agendas in global policy and research. This study examines and compares three ensemble learning models using Gradient Boosting Machine, Categorical Boosting, and Extreme Gradient Boosting for forecasting vehicle carbon dioxide (CO2) emissions. Data preprocessing with Interquartile Range (IQR) and median imputation is among the methods used to address missing values in CO₂ rating and smog rating variables. SHAP and PDP were employed for feature importance analysis and model interpretability. The findings from the third experiment demonstrate that Extreme Gradient Boosting (XGBoost) outperformed other models achieving a Coefficient Determination of 0.9988, Root-Mean-Square Error of 2.1696, Mean-Absolute Error of 0.4977, and Mean-Absolute-Percentage Error of 0.0019. The primary predictive features included combined fuel consumption (liters/100 km), city and highway fuel consumption, ethanol fuel consumption, model year, engine size and diesel consumption. The findings suggest the potential of boosting-based models for supporting sustainable transport planning, policy for emission reduction, and evidence-based policy making.

Review

The study, "Boosting-Based Machine Learning Models and Hyperparameter Tuning for Predicting Vehicle Carbon Dioxide Emission," addresses a highly relevant and pressing issue concerning sustainable development and climate change. By focusing on the prediction of vehicle CO2 emissions, the authors contribute to a critical area of environmental research with clear implications for policy and planning. The methodological approach of comparing three sophisticated ensemble learning models—Gradient Boosting Machine, Categorical Boosting, and Extreme Gradient Boosting—is well-suited to the complexity of the problem and demonstrates a robust analytical framework for achieving accurate forecasts in this domain. The technical execution of the study includes appropriate data preprocessing techniques, such as Interquartile Range (IQR) and median imputation for handling missing values, which enhances data quality and model reliability. The use of SHAP and PDP for feature importance and model interpretability is a significant strength, providing valuable insights into the drivers of CO2 emissions beyond mere predictive power. The reported performance of Extreme Gradient Boosting (XGBoost), achieving an impressive Coefficient Determination of 0.9988, alongside remarkably low error metrics, suggests a highly accurate predictive capability. The identification of key features like fuel consumption (combined, city, highway, ethanol), model year, engine size, and diesel consumption provides actionable insights for stakeholders. While the title mentions hyperparameter tuning, the abstract could benefit from briefly outlining the specific tuning strategies employed to achieve such optimal results. Overall, this research presents compelling evidence for the efficacy of boosting-based machine learning models in accurately forecasting vehicle CO2 emissions. The findings strongly support the potential for these advanced analytical tools to inform sustainable transport planning, guide emission reduction policies, and facilitate evidence-based decision-making. The high predictive accuracy of the XGBoost model, coupled with the interpretability provided by SHAP and PDP, makes this study a valuable contribution to both the machine learning and environmental policy communities, paving the way for more informed and data-driven approaches to combating climate change.

Full Text

Comments

You need to be logged in to post a comment.