Case study: fraud detection using SNA

Introduction

Humans are inherently social, and their behavior often reflects the company they keep. In the realm of fraud analytics, a principle called “homophily” signifies the likelihood of individuals having associations based on shared attributes or behaviors. A homophilic network, for instance, might comprise people from the same hometown, university, or with shared hobbies. The underlying principle is that individuals’ behavior, including fraudulent activity, might be influenced by their immediate connections. This is also sometimes referred to as “guilt by association.”

What is fraud in this context?

In the context of this case study, fraud refers to deceptive activities that may include impersonation, credit card theft, fake check submission, or any other illicit activities that can be represented and analyzed in a network of relationships. In an effort to understand the process, let’...