Innovations in Financial Crime Model Validation with Wells Fargo

Описание: Visit our website: http://bit.ly/2GtXaiw

Wayne Shoumaker, SVP Quantitative Analytics Manager, Wells Fargo

Abstract:
In financial institutions, the financial crime modeling domain is traditionally managed by experts, without much attention given to newer modeling technologies. Recent communications by regulators have encouraged exploring such technologies, and relevant examples and applications are given.

Explaining the AI/ML model
Example 1: Text classifier model
Example 2: Alert detection with tree-based approaches

ML model analysis and explainability framework
Testing above and below the line
Random forest example for Financial Crime alert detection
General framework for random forests
Sensitivity analysis for random forests

Bio:
Wayne Shoumaker is an applied statistician with experience in the financial services, healthcare, and various social services. He has been involved in basic statistical research, marketing effectiveness, business-to-business analytics, and various types of risk. His model validation experience at Wells Fargo’s Corporate Model Risk (CMoR) spans several modeling domains, such as financial crimes, fair lending, credit and operations risk, and CCAR,. He leads a team within CMoR’s Decision Science/Artificial Intelligence (DSAI) group, which validates traditional and newer technology artificial intelligence and machine learning models. Such models are used within various channels to classify customer and employee behaviors for a variety of activities. Wayne has a Ph.D. in Quantitative Psychology from the University of California at Berkeley, an M.A. in General Experimental Psychology, and a J.D. in Law.