A Risk-defined Trust Transitivity Model for Group Decisions in Social Networks

In a social network group decision making process, trust is a critical source of information. Trust transitivity is necessary in many cases to evaluate experts' trustworthiness through referring paths. However, existing trust transitivity models have hidden assumptions regarding the risk attitude of the buyer. Moreover, some users are more conservative than others regarding the risk bearing on referrals. To address this issue, this paper proposes a risk-defined trust transitivity model for group decision making in social networks with four tuple information, namely, trust, distrust, uncertainty and inconsistency. Firstly, a novel method of measuring knowledge degree is defined and applied to trust ranting. Afterwards, we develop a risk-defined trust propagation operator that can propagation trust and distrust information based on risk bearing level. To aggregate multiple possible paths, we propose a trust aggregation operator based on path centrality. Finally, we demonstrate the effectiveness of the trust model through an example of trusted service selection and verify the effectiveness of the proposed method through comparison with existing models.