MATLAB and Simulink Seminars

Model Risk Management with MATLAB


Overview

Adherence to regulations is critical, but how do you ensure compliance when they are a moving target? Traceability, reproducibility, and reusability are the three pillars upon which models should be constructed to meet regulatory requirements. Doing so will naturally reduce operational costs yet provide the agility necessary for the risk team to adapt.

A platform is necessary to address key challenges around model implementation, validation, governance, and review. Furthermore, such an environment must have many automation capabilities (documentation generation, model validation automation, etc.) and provide complete model lineage to ensure adherence to regulatory standards such as SR11-7, Basel, CCAR, CECL, and FRTB. Finally, making use of existing models and providing programming language interoperability is crucial given we now live in a polyglot world. All this information must then be summarized in customizable dashboards to be consumed on demand by key stakeholders.

In this session, we will demonstrate MATLAB’s MRM capabilities around model development (advanced AI/deep learning), model lineage, documentation automation, model review & monitoring, all tracked within MathWorks Model Inventory making MATLAB® the complete model management solution as opposed to just an ad-hoc model development tool.

Highlights

  • Using MathWorks Model Inventory to manage the model lifecycle
  • Model and documentation automation
  • Creating credit scorecards, automated predictor selection, and binning
  • Building challenger models using advanced AI and machine learning techniques
  • Interoperability with open source tools to leverage existing models
  • Providing a collaborative environment for model review
  • Powerful tools to take prototypes into production
  • Creating customizable dashboards and performing on-going model monitoring

Who Should Attend

People with an interest in:

  • Model Development
  • Model Validation
  • Model Risk Management
  • Risk Transformation

About the Presenter

Ian McKenna joined MathWorks in 2011 as an Application Engineer supporting the Financial Services industry.  During this time, his focus has been in Computational Finance with applications ranging from risk management, portfolio optimization and asset allocation, time series forecasting, to instrument pricing.  Prior to joining MathWorks he worked at the University of British Columbia developing simulation code used in industry for heat treatment of steel alloys.  Ian holds a Ph.D. from Northwestern University and a B.S. from the University of Florida in Materials Science and Engineering with a minor in Business Administration.

Paula Poza is an Applications Engineer at MathWorks, focusing on the Finance industry. She holds a degree in Mathematics from Spain and Actuarial studies from the Institute of Actuaries in UK. Her professional career prior to MathWorks was spent in consultancies and financial entities in UK and Spain.

Agenda

Time Title

10:00

Introduction

10:10

Model Inventory

  • Tracking the Model Lineage and understanding Model Dependencies
  • Customizable executive Dashboards
  • Implementation of alerts, triggers, and actions for ongoing model monitoring
10:45

Automation

  • End-to-end document and code lineage
  • Automation of documentation, status tracking, and testing
  • Elimination of refactoring and recoding models for production
11:25

Q&A

Registration closed