Detecting Frauds Using Customer Behavior Trend Analysis and Known Scenarios

The present paper proposes a fraud detection method in which user behaviors are m odelled th ough using two main components known as abnormal trend analysis component and sce nario -based component. The extent of deviation of a transction from customers’ normal behavior is estimated u sing fuzzy membership functions. The results of applyi ng all mem bership fu nctions on a transaction will then be infused, and a fin l risk is de termined as for decidin g whether to block the arrived tr ansaction orthe basis ptimized threshold f or the val e of the f inal risk is not. An estimated in order to strike a balance betw en fraud d etection rate of such problems is Although the assessment and alar m rate. useful in application m ethod is sh own to be complicated, this according to several measures and metrics.

Detecting Frauds Using Customer Behavior Trend Analysis and Known Scenarios Keywords:

fraud detection , user profiles , fuzzy membership function , trend analysis , scenario