A model governance analyst must: (1) explain to a specific loan applicant why their application was denied, and (2) produce an executive report on which features most influence the model's decisions across the entire customer portfolio. Which approach satisfies both requirements?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Think of it this way: in real-world AI governance, use shap for both requirements because shap uses shapley values to produce both local explanations for individual predictions and global feature importance across all predictions; lime only provides local explanations is exactly what teams reach for when they need to handle this scenario. SHAP (SHapley Additive exPlanations) uses Shapley values to generate both local explanations for individual predictions and global feature importance across all predictions. On the exam, remember that this falls squarely under the 1.0 AI Governance Overview domain.
Full explanation below image
Full Explanation
SHAP (SHapley Additive exPlanations) uses Shapley values to generate both local explanations for individual predictions and global feature importance across all predictions. LIME is limited to local explanations for individual predictions only. The correct answer, "Use SHAP for both requirements because SHAP uses Shapley values to produce both local explanations for individual predictions and global feature importance across all predictions; LIME only provides local explanations", directly addresses the scenario described because it aligns with the specific governance requirement in question. The incorrect options ("Use LIME for both requirements because LIME generates feature importance scores at both the individual prediction level and globally across all predictions", "Use SHAP for the individual explanation and LIME for the portfolio-level report because SHAP is computationally too intensive for single-prediction explanations", "Use LIME for the individual denial explanation and SHAP for the portfolio report because LIME provides global explanations while SHAP provides only local ones") may seem plausible but do not satisfy the core requirement. Understanding the distinction between these concepts is critical for IBM watsonx.governance implementations and is frequently tested in the 1.0 AI Governance Overview section of the certification exam.