Scholarlink

<p>The integration of artificial intelligence in economic modeling has revolutionized sustainability forecasting, yet the “black box” nature of advanced machine learning models poses significant challenges for policy transparency and regulatory compliance in economic decision-making. This study addresses the critical need for explainable AI (XAI) in industrial sustainability prediction to enhance model interpretability while maintaining predictive accuracy for economic policy formulation. We develop a comprehensive XAI framework combining machine learning excellence with transparent interpretation for industrial sustainable development forecasting. Leveraging datasets from the OECD, World Bank, IEA, and EEA covering 150 countries from 2000-2020, we systematically evaluate eight machine learning algorithms and apply SHapley Additive exPlanations (SHAP) methodology to quantify feature contributions and reveal complex economic relationships. The XGBoost model achieved superior predictive performance (accuracy = 0.881, F1-score = 0.897) in economic sustainability forecasting. Our SHAP-based explainability analysis identified green innovation (SHAP = 0.82), clean energy ratio (0.78), and ESG scores (0.77) as primary positive economic drivers, while carbon intensity (-0.75) emerged as the dominant negative factor. Scenario-based SHAP analysis revealed that high sustainability probability (89%) requires coordinated advancement across innovation and environmental dimensions, providing quantitative evidence for multi-factor economic policy design. This XAI framework transforms complex sustainability data into actionable economic insights, enabling evidence-based policy formulation while ensuring algorithmic transparency required for regulatory compliance. Our research demonstrates how explainable AI can bridge the gap between sophisticated economic forecasting models and practical policy transparency requirements, advancing the application of interpretable machine learning in economic modeling and forecasting domains.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Explainable Machine Learning for Industrial Sustainability Forecasting Using SHAP Analysis in Economic Policy Decision Support

  • 期刊论文

  • 发表日期：

    2026年3月16日

查看全文
  • Yihang Zhang,
  • Cheng Yang,
  • Chunhong Yuan

摘要

The integration of artificial intelligence in economic modeling has revolutionized sustainability forecasting, yet the “black box” nature of advanced machine learning models poses significant challenges for policy transparency and regulatory compliance in economic decision-making. This study addresses the critical need for explainable AI (XAI) in industrial sustainability prediction to enhance model interpretability while maintaining predictive accuracy for economic policy formulation. We develop a comprehensive XAI framework combining machine learning excellence with transparent interpretation for industrial sustainable development forecasting. Leveraging datasets from the OECD, World Bank, IEA, and EEA covering 150 countries from 2000-2020, we systematically evaluate eight machine learning algorithms and apply SHapley Additive exPlanations (SHAP) methodology to quantify feature contributions and reveal complex economic relationships. The XGBoost model achieved superior predictive performance (accuracy = 0.881, F1-score = 0.897) in economic sustainability forecasting. Our SHAP-based explainability analysis identified green innovation (SHAP = 0.82), clean energy ratio (0.78), and ESG scores (0.77) as primary positive economic drivers, while carbon intensity (-0.75) emerged as the dominant negative factor. Scenario-based SHAP analysis revealed that high sustainability probability (89%) requires coordinated advancement across innovation and environmental dimensions, providing quantitative evidence for multi-factor economic policy design. This XAI framework transforms complex sustainability data into actionable economic insights, enabling evidence-based policy formulation while ensuring algorithmic transparency required for regulatory compliance. Our research demonstrates how explainable AI can bridge the gap between sophisticated economic forecasting models and practical policy transparency requirements, advancing the application of interpretable machine learning in economic modeling and forecasting domains.