An Application of Unsupervised Fraud Detection to Passenger Name Records

Fraud is a threat that most online service providers must address in the development of their systems to ensure an efficient security policy and the integrity of their revenue. If rule-based systems and supervised methods usually provide the best detection and prevention, labelled training datasets are often non-existent and such solutions lack reactivity when facing adaptive fraudsters. Many generic fraud detection solutions have been made available for companies though cannot compete with dedicated internal implementations. This study presents an evaluation of some of the most widely used machine learning algorithms for unsupervised fraud detection applied to travel booking information represented by Passenger Name Records (PNR). The current paper also highlights the use of some aggregation functions relying on fuzzy logic and interpolation as an extension of unsupervised ensemble learning.