【摘要】 在线社会化媒体中意见领袖的准确测度与识别,对于重大突发舆情的情报研判、引导干预与妥善处置具有重要的研究意义。为此,首先从用户的静态网络结构特征与动态信息交互两方面入手,选择7项测度指标,基于熵权灰色关联方法,建立综合测度指标体系中的关联系数、指标权重以及加权关联度的计算模型。接下来,爬取天涯社区国际观察板块中,有关"522新疆暴恐事件"话题讨论中的用户关系与交互数据,对提出的意见领袖测度计算模型进行全面的实验分析。最终得出结论,与单一考虑某个测度指标的方法相比,提出的综合意见领袖测度模型能够更加全面、准确地识别出在线社会化媒体中的高影响力用户群体。研究成果将为我国在重大突发舆情监控、应对处置等方面提供有力的方法支持。更多还原

【Abstract】 It has been an important research issue for information analysis, government intervention, and proper disposal during major sudden public opinion spreads to accurately measure and identify opinion leaders on online social networks. Seven metrics for identifying opinion leaders are selected from two aspects of user’s static network structure characteristics and dynamic information interactions. Subsequently, based on Grey relational analysis and entropy value method, the computational model of correlation coefficient, metric weight, and weighted correlation degree in the measurement system is constructed. Finally,this paper fetches the data of users’ relationship and interactions in the topic discussions about "522 incident of violence and terrorism in Xinjiang" in the international review forum of Tianya and subsequently conducts elaborate experiments to verify the proposed computational model for discovering opinion leaders. In summary,compared with the other methods with a single metric,the proposed integrated model can more comprehensively and accurately identify opinion leaders on online social network platforms.This achievement can provide powerful methodological support to our country to monitor and handle major sudden public opinions.更多还原