Machine Learning for Climate Policy: Understanding Policy Progression in the European Green Deal

2025-10-17

AI Summary

This study addresses the problem of predicting the legislative progress of climate policies under the European Green Deal—from announcement to adoption. We construct a novel dataset comprising 165 policies, integrating textual content and structured metadata (e.g., policy initiators, countries, political parties). Methodologically, we propose ClimateBERT, a domain-adapted BERT variant, and employ multimodal joint modeling that fuses textual embeddings with heterogeneous metadata features. Results show that ClimateBERT achieves optimal performance using text alone (RMSE = 0.17, R² = 0.29); incorporating metadata further improves prediction accuracy (RMSE = 0.16, R² = 0.38), validating the efficacy of cross-modal feature synergy. Leveraging explainable AI (XAI) techniques, we systematically uncover how semantic sentiment in policy texts and political attributes jointly influence legislative trajectories—marking the first such analysis in climate policy forecasting. This significantly enhances transparency, interpretability, and predictive precision in climate governance research.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Natural Language Processing: Language Grounding & Multi-modal NLP
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Climate change demands effective legislative action to mitigate its impacts. This study explores the application of machine learning (ML) to understand the progression of climate policy from announcement to adoption, focusing on policies within the European Green Deal. We present a dataset of 165 policies, incorporating text and metadata. We aim to predict a policy's progression status, and compare text representation methods, including TF-IDF, BERT, and ClimateBERT. Metadata features are included to evaluate the impact on predictive performance. On text features alone, ClimateBERT outperforms other approaches (RMSE = 0.17, R^2 = 0.29), while BERT achieves superior performance with the addition of metadata features (RMSE = 0.16, R^2 = 0.38). Using methods from explainable AI highlights the influence of factors such as policy wording and metadata including political party and country representation. These findings underscore the potential of ML tools in supporting climate policy analysis and decision-making.

Problem

• Predicting climate policy progression status using machine learning methods
• Comparing text representation approaches for analyzing European Green Deal policies
• Evaluating metadata impact on policy adoption prediction performance

Innovation

• Using ClimateBERT to analyze policy text progression
• Combining BERT with metadata for superior prediction accuracy
• Applying explainable AI to identify key policy factors