Instance-Level Explanations for Fraud Detection: A Case Study
This addresses the challenge of interpretability in fraud detection for domain experts, but it is incremental as it applies existing explanation techniques in a new case study.
The paper tackled the problem of verifying fraud detection predictions by designing two novel dashboards that combine state-of-the-art explanation techniques, enabling domain experts to analyze predictions and dramatically speed up filtering potential fraud cases.
Fraud detection is a difficult problem that can benefit from predictive modeling. However, the verification of a prediction is challenging; for a single insurance policy, the model only provides a prediction score. We present a case study where we reflect on different instance-level model explanation techniques to aid a fraud detection team in their work. To this end, we designed two novel dashboards combining various state-of-the-art explanation techniques. These enable the domain expert to analyze and understand predictions, dramatically speeding up the process of filtering potential fraud cases. Finally, we discuss the lessons learned and outline open research issues.