Product customization strategies using the observed data of consumer behavior

摘要

A computer adapted for pay-as-you go or other metered use has a policy for determined what measurements to take to detect fraud as well as steps to take when fraud is found. To optimize between good performance and sufficient tests to reduce the risk of fraud, a policy is developed based on observation of the users behavior, using data taken at the computer, data from a payment processor or both. After analysis, an updated policy is securely loaded at the computer to determine what, and how often to measure for suspected fraud.