Fraud Networks
Data Skeptic
00:00
Tackling Class Imbalance in Fraud Detection
This chapter discusses the challenges posed by class imbalance in fraud detection, focusing on machine learning algorithms designed to identify fraudulent transactions among numerous legitimate ones. The introduction of the iFraud simulator illustrates how synthetic datasets can help researchers navigate issues related to data biases and availability, enabling more accurate modeling of fraud networks. The conversation highlights the complexities of simulating fraud detection systems, the importance of social networks, and the necessity of continuously updating predictive models to respond to evolving fraudulent tactics.
Transcript
Play full episode
