Centre for Business Analytics and Performance - Webinar - Dr. Tom Zougas

Dr Tom Zougas PEng PhD, VP Worldwide Data Analytics Consulting, Altair Canada 

Fraud detection and prevention is both an interesting and important topic because fraud costs everyone money, from the individual consumer to large corporations.

Traditional rules-based systems may have been effective in the past in identifying fraud, but they become ineffective and stale as fraudsters learn how to bypass those rules. It becomes even more challenging due to the large volumes of data that need to be processed and examined to detect fraud, in addition to the constantly changing tactics for committing fraud – those activities are usually hidden in large volumes of data.

Data-driven methods of detecting fraud leverage machine learning techniques that have been developed over the past few years and have become effective in detecting fraud with the advances in data systems (e.g. big data, streaming data) and computational systems (e.g. high performance computing, GPU). As a result, it is possible to leverage the data that is constantly being captured during our day to day activities, to identify patterns of behaviour that are associated with fraudulent activity. In addition, it is feasible to address the challenge associated with fraudsters changing their tactics.

In this presentation, we will review what fraud is and how machine learning can help detect and prevent it. We will explore fraud use cases to describe the mechanics behind applying the machine learning techniques.