Detection of bypass fraud based on speaker recognition

In telecommunication industry, fraud becomes a serious problem that affects telecommunications service providers all around the world. As a significant amount of revenue losses to fraud every year, so an efficient system to detect fraud activities is greatly required. A well-known fraud which affects GSM and PSTN service providers is Bypass fraud. It is used to avoid a charge of international calls. However, a fraud cannot be completely eradicated, an early detection of a fraud will minimize the loss in revenue. In this paper a new approach of detection system of Bypass Fraud is proposed. The proposed system based on the speaker recognition techniques to detect any fraud based on the user profiling.