medium · FRM Part 2 Current Issues
A risk function is validating a complex Gradient Boosted Machine (GBM) model used for retail credit scoring. The developers have provided SHAP (Shapley Additive Explanations) values to meet regulatory adverse-action notice requirements.
Which of the following statements best describes a structural risk-management limitation of using post-hoc SHAP attributions compared to using an inherently interpretable model?
- SHAP values are global importance metrics and cannot provide the case-by-case 'reason codes' required by fair-lending regulations.
- The use of SHAP values inherently violates the monotonicity constraints required for conceptual soundness in credit scoring.
- SHAP values cannot handle non-linear feature interactions, making them less accurate than the GBM's internal importance metrics.
- SHAP attributions are approximations of the model's behavior and may exhibit instability across different feature resamples.
Sign up free to see the explanation and track your rank →
More FRM Part 2 Current Issues practice
- According to the BCBS standard for cryptoassets, a bank's to… — What is this limit?
- What primary balance (as a % of GDP) is required to stabilize the debt-to-GDP ratio?
- SVB's management removed interest rate hedges in 2022. According to the Economic Value of
- In monitoring model stability, which metric is commonly used to quantify the shift in dist
- What does a SHAP value of zero for a specific feature in a credit scoring model imply?
- A bank provides a 'NAV facility' to a private equity fund. If the fund's underlying assets
- What is the bank's primary ongoing obligation to regulators regarding this new model?
- If the firm's EBITDA declines, when can the lender intervene?