基于共现分析的网络议程设置研究从认知心理学的角度出发，分析媒体议程与公 众议程网络，更加贴近回答李普曼所说的“我们脑海中的图景”，但该方法只能体现 议程网络中关键词间的显性关系。本文提出通过基于上下文语义的word2vec模型，测 度网络议程设置中关键词隐性关系的研究方法。以“红黄蓝事件”为例，揭示党媒、 都市类媒体、商业媒体、知乎意见领袖以及公众的议程网络及其相似度。研究发现， 知乎意见领袖对公众议程的影响程度最大，其议程主要是对事件背后因果关系的梳理 以及对策建议的提供。在媒体方面，党媒和都市类媒体报道基本一致，且二者对公众 的影响均高于商业媒体。本研究是对现有网络议程设置研究的补充，对议程设置的本 土化研究有一定的启示。

Network agenda setting study based on co-occurrence analysis has analyzed the correlation of media agenda network and public agenda network from the perspective of cognitive psychology which expounds more closely to Lippmann’s “pictures in our heads”. However, this kind of approach can only reflect the explicit relationship between keywords in agenda networks. This study provides a research method for measuring the implicit relationship of keywords in network agenda setting through the word2vec model based on semantic context. It reveals the agenda network and its similarity of state media, metropolitan media, business media, Zhihu opinion leaders and the public through the case study of RYB Education Abuse Scandal. The study shows that Zhihu opinion leaders’ in?uence on the public is of the maximum extent of which the agenda mainly sorts out the causality behind the event and offers relevant countermeasures and suggestions. On the media side, the reports on party media and metropolitan media are basically consistent and the public in?uence of both media is respectively greater than business media. This study supplements the current researches regarding network agenda setting and may also provide certain insights to the localization study of agenda setting.