A personalized bidirectional feedback mechanism by combining cooperation and trust to improve group consensus in social network

Abstract

In the context of social network group decision making, it is an essential step to exploit the feedback mechanism for consensus reaching process (CRP) to obtain group-supported decision results. The CRP typically involves both concordant and discordant decision-makers (DMs), and the bidirectional interaction between them is regarded as a preferable approach to promote consensus. However, the existing bidirectional feedback models cannot provide personalized recommendation advice for DMs without objectively considering the DMs’ trust relationship during the interaction process. To address this issue, this paper utilizes the Uninorm propagation operator to establish the trust relationship among group DMs, which is used to generate personalized recommendation advices to modify their preference. In addition, it defines the concept of the cooperation degree index (CDI) to explore the ‘cooperative behavior’ of DMs during interaction, and then it is able to prevent over adjusting behavior. Therefore, this article establishes a novel personalized bidirectional feedback mechanism, which can reduce the opinion loss of individuals while enhancing fairness in the CRP. Finally, through case studies and discussions, the effectiveness and feasibility of the proposed approach is validated.