Grey relational analysis- and clustering-based opinion dynamics model in social network group decision making

Abstract

In combination with social networks, opinion dynamics models can effectively describe the interaction and exchange of opinions between experts in social network group decision-making (SNGDM). Usually, the opinion dynamics model allows all experts to be involved in the interaction process simultaneously. However, if a subgroup has a dominant influence in SNGDM, the authoritative subgroup can easily control the group opinion. Therefore, we propose the grey relational analysis (GRA)- and clustering-based opinion dynamics model to avoid this situation. First, this paper applies a new operator to social network analysis (SNA) to obtain complete social relationships among experts. Next, we propose an improved k-means clustering method based on experts’ social relationships and opinion similarity. Then, based on SNA, GRA, and clustering results, a novel opinion dynamics model is proposed to weaken the influence of the authoritative subgroup. Finally, the validity of the proposed model is demonstrated through numerical examples and comparative analyses.