Data Mining Techniques Applying To The Telecommunications Consumer Fraud System

The amount of fraud grows increasingly in the telecommunication industrial while the telecommunication market has made a rapid development. The anti-fraud is to control the behaviors of fraud persons, from the point of operation of telecom, according to the analysis of customer actions on the base of data mining, data mart and other advanced telecommunication technology. A Telecommunication Anti-fraud System has been built through the analysis of the specific customers associated with customer behavior analysis, customer analysis and customer credit risk analysis.At first, this paper analyzes the current background of telecommunication consumer fraud, explains the necessary and urgency of the development of a Telecommunication Anti-fraud System; then studies the technical shortcomings of the present data mining and brings out a consideration of data mining techniques applied to the Telecommunication Anti-fraud System. At second, the relative technologies of the development of a Telecommunications Anti-fraud System are stated such as data mining, data warehouse, data mart; Bayesian classifier technology. At last, the modeling techniques using data mining, Bayesian classifier and their certification processes are focused to develop. The performance test and the use in practice prove that it is market valuable to apply Bayesian classification to a Telecommunication Anti-fraud System. The system can tap the potential risks and identify customer fraud so that to solve many small-scale, decentralized telecommunications fraud.