Detecting frauds using customer behavior trend analysis and known scenarios

In this paper a fraud detection method is proposed which user behaviors are modeled using two main components namely the un-normal trend analysis component and scenario based component. The extent of deviation of a transaction from his/her normal behavior is estimated using fuzzy membership functions. The results of applying all membership functions on a transaction will then be infused and a final risk is gained which is the basis for decision making in order to block the arrived transaction or not. An optimized threshold for the value of the final risk is estimated in order to make a balance between the fraud detection rate and alarm rate. Although the assessment of such problems are complicated, we show that this method can be useful in application according to several measures and metrics.