Detection of Human and Computer Voice Spammers Using Hidden Markov Model in Voice over Internet Protocol Network

摘要

This work presents a mathematical model for Voice over Internet Protocol (VoIP) communication which detects voice spammers (both human and computer) before reaching the conversation phase. The call pattern of the voice spammers are measured which are the fundamental characteristics for detection. The proposed Hidden Markov Model (HMM) for voice spammers are compared with existing behavior-based model. Through simulations with various parameter settings, the proposed model effectively distinguishes the voice spammers from legitimate callers and prevents them before making a next call.