Recommending Users in Social Networks by Integrating Local and Global Reputation

摘要: A central research theme in the Online Social Network (OSN) scenario consists of predicting trustworthiness a user should assign to other OSN members. Past approaches predict trust relied on global reputation models: they were based feedbacks about actions performed by past and provided for entire OSN. These models have shown an evident limitation considering effects malicious fraudulent behaviors, thus making unreliable themselves. In this paper, we propose integrate with local reputation, computed ego-network. Some experiments, real datasets show that is useful only if size ego-network small, as newcomer. Besides, integrated usage reputations leads expected very high level precision.