[Construction and Analysis of Machine Learning Based Transportation Carbon Emission Prediction Model]

Area of Science:

• Environmental Science
• Data Science
• Transportation Engineering

Context:

• Carbon emissions from the transportation sector pose a significant environmental challenge.
• China's transportation sector has experienced substantial growth, leading to increased carbon footprint.
• Understanding provincial variations is crucial for effective emission reduction strategies.

Purpose:

• To construct and evaluate machine learning models for predicting transportation carbon emissions.
• To identify key factors influencing transportation carbon emissions across Chinese provinces.
• To determine the optimal machine learning algorithm for accurate emission prediction.

Summary:

• Panel data from 30 Chinese provinces (2005-2019) were used to train and test seven machine learning algorithms.
• XGBoost and Random Forest models demonstrated high accuracy (R² > 0.97, error < 10%).
• Provincial differences, social goods consumption, and car ownership were identified as key influencing factors.

Impact:

• Provides a robust predictive framework for transportation carbon emissions.
• Offers data-driven insights for policymakers to design targeted emission reduction strategies.
• Highlights the importance of considering regional heterogeneity in sustainable transportation development.