A Trustworthy AI Explainability Metric for Machine Learning Volatility Models: Evidence from the EU ETS

Key Points

• AI explainability metrics enhance understanding of machine learning models, leading to improved trustworthiness.
• The metric showed a significant correlation with machine learning volatility models in EU ETS scenarios.
• Assessment using the new explainability metric evaluates performance and reliability of machine learning systems.
• Implications suggest this metric may play a central role in future regulatory frameworks for emissions trading.