摘要 |
A system compares a first sampled voice of a user with a subsequently sampled voice of a user of an identical identifier. A profile of user usage that is representative of the total number of different users of the associated identifier is determined. The profile of user usage is compared with a threshold for fraudulent use 208 and a fraud condition signal is generated in the event that the profile of user usage exceeds the threshold 214. If required, the fraudulent transaction is terminated 216. In a preferred embodiment, a fraud detection system is provided to provide a first level analysis of transactions, and to select only certain ones of the network transactions for voice sampling, which are those suspected of having a fraudulent characteristic.
|