STEMM Institute Press
Science, Technology, Engineering, Management and Medicine
Explainable Machine Learning Reveals Stage-Dependent Predictors of Global PM2.5 Exposure for Environmental Governance
DOI: https://doi.org/10.62517/jse.202611506
Author(s)
Qifeng Li*, Shuo Wen
Affiliation(s)
Guangdong Police College, Guangzhou, China *Corresponding Author
Abstract
We used panel data from the World Bank covering 200 economies from 1990 to 2023. With these data, we built population-weighted PM2.5 exposure models using random forest, XGBoost, LightGBM, and linear regression. The random forest model (R² = 0.827) outperformed XGBoost (R² = 0.805), LightGBM (R² = 0.794), and linear regression (R² = 0.419). SHAP analysis showed that population size and GDP per capita made the largest contributions, followed by the share of renewable energy, forest cover, and energy consumption per capita. The marginal contribution of GDP per capita to PM2.5 falls from positive to negative across development stages, which we read as consistent with the environmental Kuznets curve hypothesis. We also observe a sign reversal between the 40%–60% and 60%–80% urbanization bands. Population, renewable energy, and forest cover show stage-specific reversals as well. Our results suggest that economic growth and PM2.5 are not monotonically coupled, and that the link between structural conditions and exposure levels varies across development stages. We believe these findings provide data-driven and interpretable evidence for phased air quality management and sustainable development planning. They also reveal where environment-related public complaints are most likely to accumulate, which gives an evidence base for allocating social governance resources and preventing public order risks.
Keywords
PM2.5; Machine Learning; SHAP; Environmental Kuznets Curve; Sustainable Development; Socioeconomic Correlates
References
[1] World Health Organization. WHO Global Air Quality Guidelines: Particulate Matter (PM2.5 and PM10), Ozone, Nitrogen Dioxide, Sulfur Dioxide and Carbon Monoxide; World Health Organization: Geneva, Switzerland, 2021. [2] Cohen, A.J.; Brauer, M.; Burnett, R.; Anderson, H.R.; Frostad, J.; Estep, K.; et al. Estimates and 25-Year Trends of the Global Burden of Disease Attributable to Ambient Air Pollution: An Analysis of Data from the Global Burden of Diseases Study 2015. Lancet 2017, 389, 1907–1918. [3] Burnett, R.; Chen, H.; Szyszkowicz, M.; Fann, N.; Hubbell, B.; Pope, C.A., III; et al. Global Estimates of Mortality Associated with Long-Term Exposure to Outdoor Fine Particulate Matter. Proc. Natl. Acad. Sci. USA 2018, 115, 9592–9597. [4] Shaddick, G.; Thomas, M.L.; Mudu, P.; Ruggeri, G.; Gumy, S. Half the World's Population Are Exposed to Increasing Air Pollution. npj Clim. Atmos. Sci. 2020, 3, 23. [5] Grossman, G.M.; Krueger, A.B. Economic Growth and the Environment. Q. J. Econ. 1995, 110, 353–377. [6] Shafik, N. Economic Development and Environmental Quality: An Econometric Analysis. Oxf. Econ. Pap. 1994, 46, 757–773. [7] Selden, T.M.; Song, D. Environmental Quality and Development: Is There a Kuznets Curve for Air Pollution Emissions? J. Environ. Econ. Manag. 1994, 27, 147–162. [8] Holtz-Eakin, D.; Selden, T.M. Stoking the Fires? CO2 Emissions and Economic Growth. J. Public Econ. 1995, 57, 85–101. [9] Stern, D.I. The Rise and Fall of the Environmental Kuznets Curve. World Dev. 2004, 32, 1419–1439. [10] Dinda, S. Environmental Kuznets Curve Hypothesis: A Survey. Ecol. Econ. 2004, 49, 431–455. [11] Harbaugh, W.T.; Levinson, A.; Wilson, D.M. Reexamining the Empirical Evidence for an Environmental Kuznets Curve. Rev. Econ. Stat. 2002, 84, 541–551. [12] Carson, R.T. The Environmental Kuznets Curve: Seeking Empirical Regularity and Theoretical Structure. Rev. Environ. Econ. Policy 2010, 4, 3–23. [13] Kaika, D.; Zervas, E. The Environmental Kuznets Curve (EKC) Theory—Part A: Concept, Causes and the CO2 Emissions Case. Energy Policy 2013, 62, 1392–1402. [14] Breiman, L. Random Forests. Mach. Learn. 2001, 45, 5–32. [15] Chen, T.; Guestrin, C. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; pp. 785–794. [16] Ke, G.; Meng, Q.; Finley, T.; Wang, T.; Chen, W.; Ma, W.; Ye, Q.; Liu, T.-Y. LightGBM: A Highly Efficient Gradient Boosting Decision Tree. In Advances in Neural Information Processing Systems 30; Curran Associates, Inc.: Red Hook, NY, USA, 2017; pp. 3149–3157. [17] World Bank. World Development Indicators; World Bank: Washington, DC, USA, 2025. [18] Tibshirani, R. Regression Shrinkage and Selection via the Lasso. J. R. Stat. Soc. Ser. B 1996, 58, 267–288. [19] Akiba, T.; Sano, S.; Yanase, T.; Ohta, T.; Koyama, M. Optuna: A Next-Generation Hyperparameter Optimization Framework. In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, Anchorage, AK, USA, 4–8 August 2019; pp. 2623–2631. [20] Lundberg, S.M.; Lee, S.-I. A Unified Approach to Interpreting Model Predictions. In Advances in Neural Information Processing Systems 30; Curran Associates, Inc.: Red Hook, NY, USA, 2017; pp. 4765–4774. [21] Lundberg, S.M.; Erion, G.; Chen, H.; DeGrave, A.; Prutkin, J.M.; Nair, B.; Katz, R.; Himmelfarb, J.; Bansal, N.; Lee, S.-I. From Local Explanations to Global Understanding with Explainable AI for Trees. Nat. Mach. Intell. 2020, 2, 56–67. [22] Štrumbelj, E.; Kononenko, I. Explaining Prediction Models and Individual Predictions with Feature Contributions. Knowl. Inf. Syst. 2014, 41, 647–665.
Copyright @ 2020-2035 STEMM Institute Press All Rights Reserved