Abstract: A mechanism is provided for dynamically evaluating fraud risk. Responsive to receiving a notification, a set of indicators associated with a focal object are automatically analyzed to form a combined assessment. The combined assessment is compared to a set of pattern fraud determinations. The combined assessment is then adjusted based on the set of pattern fraud determinations to dynamically determine a current fraud assessment of the focal object. An action is them performed based on the current fraud assessment.

Abstract: A computer-implemented method, system, and computer program product for optimizing analytics selection to detect fraud retrieves, from an electronic database, historical data on a plurality of available fraud detection analytics, the historical data including service times and risk levels associated with the fraud detection analytics, and analyzes the historical data to form a fraud detection analytics optimization model. A required service time for the transaction is determined based on information about the requester and the transaction, and the fraud detection analytics optimization model is used to select, from a plurality of available fraud detection analytics, a fraud detection analytic with a lowest risk level predicted to execute within the required service time.