Abstract:Trust evaluation in a network is important in many areas, such as group decision-making and recommendation in e-commerce. Hence, researchers have proposed various trust network models, in which each agent rates the trustworthiness of others. Most of the existing work require the agents to provide accurate degrees of trust and distrust in advance. However, humans usually hesitate to choose one among several values to assess the trust in another person and tend to express the trust through linguistic descriptions. Hence, this paper proposes a novel trust network model that takes linguistic expression of trust into consideration. More specifically, we structure trust scores based on hesitant fuzzy linguistic term sets and give a comparison method. Moreover, we propose a trust propagation method based on the concept of computing with words to deal with trust relationships between indirectly connected agents, and such a method satisfies some intuitive properties of trust propagation. Finally, we confirm the advantages of our model by comparing it with related work.

A Trust Network Model Based on Hesitant Fuzzy Linguistic Term Sets