Course Objective & Scope
In finance, insurance and other regulated sectors, no model should go live without an independent review. The programme covers model inventory and risk tiering, conceptual soundness review, data quality tests, out-of-time backtesting, stability analysis with PSI and CSI, challenger model comparison, and sensitivity and stress testing.
SR 11-7 style model risk management principles are discussed alongside the testing and documentation expectations of the EU AI Act for high-risk systems. Participants prepare a validation report with graded findings on a sample model.
Who Should Attend?
- Model validation and model risk management teams
- Data scientists building credit, insurance and pricing models
- Technical specialists in internal audit and compliance
- Prerequisite: experience developing or reviewing statistical models
Key Takeaways
- Set validation scope and depth according to model risk
- Design out-of-time and out-of-sample backtests
- Monitor stability with PSI, CSI and discriminatory power metrics
- Challenge the developer's model independently with challenger models
- Document findings with severity ratings in a validation report
Curriculum
Model Risk Management Framework
- Model inventory, risk tiering and three lines of defence
- SR 11-7 principles and EU AI Act testing requirements
Conceptual Soundness and Data Review
- Challenging assumptions, variable selection and target definition
- Data quality, representativeness and leakage checks
Performance and Backtesting
- Gini, KS, AUC and calibration tests
- Out-of-time backtesting and period comparisons
Stability, Sensitivity and Stress Testing
- Population shift analysis with PSI and CSI
- Scenario-based stress and input sensitivity
Challenger Models and the Validation Report
- Building an independent benchmark model
- Finding severity, action plans and monitoring recommendations
Other Courses in This Category
-
Time series forecasting with ARIMA, ETS, Prophet and LightGBM-based approaches, working hands-on with seasonality, exogenous variables, hierarchical series, backtesting and prediction intervals to build reliable forecasts.
- Live Course
- 5 modules
-
Detecting unusual behaviour in unlabelled data with statistical methods, Isolation Forest, LOF, autoencoders and time series anomaly techniques, including threshold setting, alert fatigue and evaluation strategies.
- Live Course
- 5 modules
-
Applied Machine Learning
End-to-end machine learning with Scikit-learn, XGBoost, LightGBM, and CatBoost, covering feature engineering, Optuna hyperparameter tuning, imbalanced data, and SHAP/LIME explainability.
- Live Course
- 3 modules
-
Behavioural features, gradient boosting, anomaly scores and hybrid rule-model setups for payment, insurance and account fraud, with hands-on work on extreme class imbalance, delayed labels and operational threshold management.
- Live Course
- 5 modules
-
Variable analysis with WoE/IV, logistic regression scorecards, gradient boosting alternatives, reject inference and score calibration for credit, insurance and supplier risk, building explainable and auditable risk scores.
- Live Course
- 5 modules
-
Churn, purchase and cross-sell propensity models covering target definition, customer behaviour features, survival analysis, uplift modelling and campaign targeting, turning predictions into measurable retention and sales actions.
- Live Course
- 5 modules
Get in touch
