Machine Learning for Fraud Prevention

In an era when eCommerce is experiencing an accelerated digital transformation, fraud practitioners are increasingly embracing digital tools and solutions to aid in the fight against evolving fraud. In this course, fraud and risk specialists will learn the fundamentals of machine learning (ML) and how they can leverage ML to help identify and prevent fraud in a card-not-present setting

The course is developed by the MRC in partnership with Signifyd and peer-reviewed by the MRC Education Committee.

Program Details

• Program Level: Basic
• Program Field of Study: Specialized Knowledge
• Program Delivery Method: QAS Self Study
• CPE Credits: 2.5
• Advanced Preparation and/or Pre-requisites: None

Objectives

• Define machine learning
• Distinguish types of models and their applications in fraud management
• Compare and contrast the advantages and disadvantages between the traditional approach to fraud fighting and machine learning
• Convert data into features that a model can use
• List methods to label transactions as fraudulent or not
• Deconstruct the components of a model
• Identify key metrics for evaluating machine learning models
• Examine common types of drift that can impact a model’s performance
• Define model scores
• Differentiate between raw scores and calibrated scores
• Explain decision thresholds
• Summarize how thresholds can be combined with rules and manual review to make decisions

The Machine Learning for Fraud Prevention course is priced at US$150 for MRC members (plus applicable taxes). Non-members may take the course at the non-member price of US$300 (plus applicable taxes).