Related Experiment Video
Updated: Jul 24, 2026
09:34
Implementation of Portable Emissions Measurement Systems PEMS for the Real-driving Emissions RDE Regulation in Europe
Published on: December 4, 2016
28.1K
A machine learning algorithm to explore the drivers of carbon emissions in Chinese cities
Wenmei Yu1, Lina Xia1, Qiang Cao2
1School of Finance, Anhui University of Finance and Economics, Bengbu, 233030, China.
Scientific Reports
|October 9, 2024
Summary
Energy consumption drives China
Area of Science:
- Environmental Science and Policy
- Applied Economics
- Data Science
Background:
- China's role as a major energy consumer and carbon emitter necessitates effective carbon emission reduction strategies.
- Understanding the complex, non-linear relationships between economic factors and carbon emissions is crucial for achieving dual-carbon goals.
- Existing econometric models may not fully capture the nuances of these relationships.
Purpose of the Study:
- To compare the predictive performance of six machine learning algorithms against traditional econometric models for China's carbon emissions.
- To identify and analyze the key domestic economic, external economic, and policy uncertainty factors influencing carbon emissions.
- To investigate the non-linear relationships between these factors and carbon emissions using the Extra-trees model.
Main Methods:
- Utilized panel data from 254 Chinese cities spanning 2011-2020.
- Employed and compared six distinct machine learning algorithms.
- Applied the Extra-trees model to determine factor importance and analyze non-linear relationships via Partial Dependence Plots (PDPs).
Main Results:
- Energy consumption (ENC) is the primary driver of increased carbon emissions; government intervention (GOV) and digital finance (DIG) show significant reduction effects.
- Foreign direct investment (FDI) and economic policy uncertainty (EPU) significantly influence carbon emissions, with PDPs supporting the pollution haven hypothesis and indicating EPU's emission-reducing role.
- Energy consumption (ENC) is a universal driver across different city sizes, though variations exist.
Conclusions:
- Machine learning models offer valuable insights into carbon emission drivers beyond traditional methods.
- Policy interventions targeting energy consumption, promoting digital finance, and managing economic policy uncertainty are key for emission reduction.
- Tailored strategies considering city-specific characteristics are essential for sustainable development pathways.
