Optimizing Fraud Detection Using Machine Learning Techniques

• Activity Number: 427 - Intelligent Systems and Decision Support
• Type: Contributed
• Date/Time: Wednesday, August 10, 2022 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Jennifer Renee Leach and Umashanger Thayasivam*
• Companies: Rowan University and Rowan University
• Keywords: machine learning; fraud detection; supervised learning

Abstract

As technology has become a more prominent and important part of our society, it has also brought along consequences with it. With the advancement of technology comes advancements within the system of fraud. People are constantly learning new ways to deceive people for their own personal gain, and because of this, it is important we are constantly finding new ways to detect these fraudulent actions. The machine learning techniques used within the field of data science pose a potential solution to this issue. These different techniques allow us to create models that can later be used to help detect fraud before, or as, it is happening. We explore the use of various machine learning techniques across multiple types of fraud in order to find the method which produces the optimal accuracy, recall, precision, and F1.