Yuhas, B. P. (1995, April). Learning the structure of telecommunications fraud. Technical report, Bellcore.  Home/Search   Document Not in Database   Summary   Related Articles

....of fraud episodes in order to generate features (measures) that distinguish fraudulent from legitimate behavior. To the best of our knowledge, no published anomaly detection system does this. Calling card fraud and credit card fraud are other forms of superimposition fraud. A system built by Yuhas (1993, 1995) examines a set of records representing callingcard validation queries to identify queries corresponding to fraudulent card usage. Yuhas transformed the problem into a two class discrimination task and trained several machine learning models on the data. All three models had comparable 24 ....

....Hidden Markov Models (Rabiner and Juang 1986; Smyth 1994) are concerned with distinguishing recurring sequences of states and the transitions between them. However, fraud detection usually only deals with two states (the frauded and un frauded states) with a single transition between them. Yuhas (1995) mentions the possibility of recognizing at home and travel states in order to distinguish frauded states more effectively. This differentiation could be useful for reducing false alarms. We are aware of no work pursuing this idea. 8. Conclusion The detection of cellular cloning fraud is a ....

Yuhas, B. P. (1995, April). Learning the structure of telecommunications fraud. Technical report, Bellcore.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC