Financial Innovation

Table 1 Components of ML model risk

From: Measuring the model risk-adjusted performance of machine learning algorithms in credit default prediction

Model risk components
Statistics
Stability
Hyper-parameters
Over-fitting
Dynamic calibration
Feature engineering
Technology
Transparency
Carbon-footprint
Third-party providers
Cyber-risk
Market conduct
Privacy
Auditability
Interpretability
Biases

Back to article page