SPIT Detection in Telephony over IP Using K-Means Algorithm

Abstract

In few recently years, the number of unwanted messages and calls through IP telephony are increased rapidly. These undesirable messages and calls are booth annoy the receiver. However, the unsolicited calls are more annoying the callee, because he has to response to the call in real-time. SPIT tries to bring a loss in financial and to make ToIP non preferable for costumer. Thus, detection of SPIT is an indispensable challenge to solve if we want to enhance the quality of services in ToIP network. In this work, we give more details about previous proposed approaches to detect SPIT. Then, we present our proposed approach which is based on calculation the value of six features. In our work, we use K-means algorithm to cluster the caller as spammer or non-spammer. The experimental results show that our proposed approach can achieve True Positive Rate between 83.3% and 99.23%, depending on the number of SPIT in the network.