DALLAS-Northern Telecom Ltd. entered the fraud-detection and prevention market with its new product offering called SuperSleuth, which combines neural network-based technologies with more traditional rules-based detection techniques to help prevent false fraud alarms.
Not unlike other fraud-prevention systems, SuperSleuth establishes individual calling profiles for subscribers and analyzes customer data records in near real-time for evidence of fraudulent activity, such as multiple overlapping calls from the same phone number and calls of an extremely long duration.
SuperSleuth will alert a carrier’s fraud analysts when apparent fraudulent calling patterns are detected, regardless of its switch or network technology.
SuperSleuth’s neural network-based subscriber profiler goes further than traditional rules-based software to actually mimic the learning patterns of the human brain, differentiating between apparent fraudulent use and seemingly similar but valid phone use, said Nortel.
The adaptive approach allows the system to learn continually and respond to the characteristics of known patterns of fraud, as well as new kinds of fraud that could be introduced on the network.
“Neural networks learn quickly, adapt to human behavior and are a highly effective method in identifying patterns which may indicate fraudulent calls,” said Bill Seymour, general manager of Fraud Solutions for Nortel.
“Calling patterns change rapidly in the wireless industry, and the need for a tool with the abilities of SuperSleuth is a necessity in today’s competitive environment where revenue stream is key.”
Neural networks convert data into meaningful information, which allows millions of calls to be grouped into behavioral profiles for individual subscribers, resulting in the generation of far fewer false alarms. The system’s neural network-based capability is complemented by a front-line defense using rules-based technology to monitor calls for a number of red flags that could indicate potential fraud is underway.
“SuperSleuth allows fraud analysts to significantly increase their productivity, allowing operators of wireless networks to save time and money,” said Seymour.
