Outlier detection and its applications in the fraud detection

Abstract

Fraud detection is an important task for many organizations is today’s connected and rapidly changing world. The use of an outlier detection method is a common way of dealing with fraud detection. Numerous outlier detection techniques have been developed and researched within diverse research and application domains. In this paper, we try to present a comprehensive overview of different outlier detection methods and applications in the fraud detection. The choice of an appropriate method is important, therefore we identified some possible factors that can influence the method choice. Different methods for outlier detection are provided and structured by grouping them into categories. A basic explanation and some examples are given for each method, as well as advantages and disadvantages per category. Furthermore, we collected common fraud detection applications and analyzed how the chosen outlier detection methods handle specific outliers. We hope this paper provides a better understanding of the possible directions and challenges of outlier detection methods and their uses in the fraud detection domain.