The client needed to test and validate its Identity Fraud Model for the Identity Theft Prevention & Red Flags Program.

Solytics performed testing and validation of the Identity Fraud Model:

1. Reviewed all model validation and testing documents for completeness and accuracy

2. Selected the Gradient Boosting Model (GBM) as the Benchmark Model and Logistic Regression, ANN, Random Forest, SVM, etc as the Challenger models.

3. Tested the Benchmark Model against the Challenger models to evaluate model performance.

4. Suggested additional tests and conducted regular discussions with stakeholders.

5. Provided insights for future model use and developed thorough validation documents.

- Finalized and implemented the GBM model based on it outperforming all challenger models.

- Enhanced the quality of model documentation.

- Improved overall model performance metrics.

- Provided a comprehensive testing and validation document to be used for CCAR and other regulatory submissions.