电子科技大学学报
電子科技大學學報
전자과기대학학보
Journal of University of Electronic Science and Technology of China
2015年
5期
749-756
,共8页
信任边界%复杂网络%连续观点%观点演化
信任邊界%複雜網絡%連續觀點%觀點縯化
신임변계%복잡망락%련속관점%관점연화
bounded confidence%complex network%continuous opinion%opinion evolution
多数的连续观点演化模型中,缺少考虑舆论环境对信任关系的影响,同时信任边界条件使得观点演化聚合之后,各观点集合中的个体无法进行后续观点交互。该文分析个体观点差异以及邻居舆论环境对交互过程中信任关系的影响,消除信任边界对观点交互更新的限制,建立无观点信任边界限制的连续观点演化模型,并研究该模型观点演化统一的影响机制。理论分析与仿真实验表明:该模型不单可以反映观点演化的聚合过程,而且能反映舆论演化聚合之后各观点集合之间的竞争交互过程,模型可描述观点集合形成后的观点有序集合的变化;同时,模型反映了在较高聚类系数的无标度网络中,观点演化统一更加困难,而在平均连接度高的网络中,演化统一更加容易,这与已有相关研究结论是基本相符的。
多數的連續觀點縯化模型中,缺少攷慮輿論環境對信任關繫的影響,同時信任邊界條件使得觀點縯化聚閤之後,各觀點集閤中的箇體無法進行後續觀點交互。該文分析箇體觀點差異以及鄰居輿論環境對交互過程中信任關繫的影響,消除信任邊界對觀點交互更新的限製,建立無觀點信任邊界限製的連續觀點縯化模型,併研究該模型觀點縯化統一的影響機製。理論分析與倣真實驗錶明:該模型不單可以反映觀點縯化的聚閤過程,而且能反映輿論縯化聚閤之後各觀點集閤之間的競爭交互過程,模型可描述觀點集閤形成後的觀點有序集閤的變化;同時,模型反映瞭在較高聚類繫數的無標度網絡中,觀點縯化統一更加睏難,而在平均連接度高的網絡中,縯化統一更加容易,這與已有相關研究結論是基本相符的。
다수적련속관점연화모형중,결소고필여론배경대신임관계적영향,동시신임변계조건사득관점연화취합지후,각관점집합중적개체무법진행후속관점교호。해문분석개체관점차이이급린거여론배경대교호과정중신임관계적영향,소제신임변계대관점교호경신적한제,건립무관점신임변계한제적련속관점연화모형,병연구해모형관점연화통일적영향궤제。이론분석여방진실험표명:해모형불단가이반영관점연화적취합과정,이차능반영여론연화취합지후각관점집합지간적경쟁교호과정,모형가묘술관점집합형성후적관점유서집합적변화;동시,모형반영료재교고취류계수적무표도망락중,관점연화통일경가곤난,이재평균련접도고적망락중,연화통일경가용역,저여이유상관연구결론시기본상부적。
Most models about continuous opinion dynamics lack the considerations of the effects of public opinion environment on trust relationships, and when opinions have been clustered into some collections, the bounded confidence will impede the subsequent opinion interactions between individuals that come from different opinion collections. In this paper, the change mechanism of trust relationships which affected by difference of opinion and neighboring opinion environment is analyzed, the restrictions of bounded confidence on individual interactions is eliminated, and a general model for the evolution of continuous opinions without bounded confidence is proposed. The results of simulation and of analysis show that our model not only reflect the opinions polymerization process, but also reflect the subsequent opinion interactions between individuals in different opinion collections after opinions have been aggregated. In addition, our model can do better at reflecting and explaining the process of opinion evolution, and it is consistent with existing search conclusions that the opinions consensus probability will be reduce if the clustering coefficient of complex network has been increased and will be increased with the addition of average degree.