Skip to main content

Open Access | Accepted manuscript on September 11, 2026

Interpretable Machine Learning for Evaluating Industry–Ecology Coupling in the Beijing–Tianjin–Hebei Region

Shi Qinyu
Mohd Yusoff Mariney
Adeline Adura Tengku Hamzah Tengku
Mohd Noor Nisfariza
Li Xiaoya
Abstract

The transformation of industrial structures and growing environmental pressures have created challenges for coordinated regional development. As a major urban agglomeration in China, the Beijing–Tianjin–Hebei region faces disparities in economic development and ecological conditions, making the identification of coordination patterns and underlying drivers important. This study investigates the spatiotemporal evolution and driving mechanisms of the coupling coordination degree between industrial structure and ecological environment across cities in the BTH region during 2010–2023. A city-level CCD model is combined with XGBoost–SHAP analysis to quantify key influencing factors’ effects and heterogeneity. The results demonstrate considerable spatial differentiation and nonlinear temporal dynamics. Beijing and Tianjin maintain relatively high coordination levels, while most cities in Hebei remain in lower coordination stages. Despite overall improvement after 2015, coordination fluctuates substantially, characterized by alternating declines and recoveries. The SHAP analysis further reveals heterogeneous effects of industrial and ecological factors, with temporal dynamic features exerting the strongest influence, indicating path dependence. Based on the resulting feature profiles, cities are categorized into four coordination types with distinct development trajectories. These findings highlight the differentiated nature of industrial–ecological coordination and underscore the need for place-specific strategies to promote sustainable and balanced regional development.

Graphical abstract
Download graphical abstract
Keywords
coupling coordination degrees, XGBoost, SHAP, driving mechanisms, spatial heterogeneity, Spatiotemporal Analysis, Sustainable Development Goals