Coevolution of opinions and actions on complex networks under bounded confidence rules

In certain specific opinion interaction scenarios, some agents express their opinions through actions or euphemisms owing to social pressure, sensitive topics, political divisions, or other factors. Motivated by Martins's seminal experiment on continuous opinions and discrete actions (CODA) in opinion dynamics, we propose a hybrid model combined with the Hegsel-mann-Krause (HK) variant model and the CODA model for studying the opinion evolution process of euphemistic agents with stubbornness and leadership in a scale-free network under homogeneous or heterogeneous situations. With this model, opinion leaders are affected by other opinion leaders instead of followers. However, followers are affected by both opinion leaders and other followers. We conclude that increasing the proportion of opinion leaders and confidence levels will promote the convergence of group opinions. In particular, when the confidence level is greater than 0.5, increasing the proportion of leaders will promote the reversal of group opinions and actions, thereby providing effective theoretical support for relevant management departments to deal with major public opinions.