社交媒体网站为网民们提供了在线交流的平台, 在该平台下各种话题 的发言层出不穷, 获取其中的热点事件有助于了解网民们关注的重点, 挖掘潜在 信息. 文章以天涯杂谈``红会贴"为数据源, 基于社会网络分析方法, 获悉了``红会贴"热点事件、网民在线回复规律, 研究表明地震捐款问题、郭美美事件等是舆情焦点, 网民双向互动的情感极性基本是相同的, 互动中先发表负面言论的数目多于先发表正面言论的数目. 由于事件的多样性、动态性, 可视化表达热点事件的演化过程尤为重要. 文章提出了结合LDA主题模型和Bert向量抽取``红会贴"故事线的方法, 由此得到``红会贴"热点事件的发展脉络, 故事线披露出了红会的信任危机.

Social media websites provide online communication platform for netizens. There are a large number of speeches on various topics on the platform. Obtaining hot events from social media can help to understand the focus of netizens' attention and mine potential information. Based on the Red Cross posts from Tianya Club, this paper uses the social network analysis method to learn hot events of the Red Cross posts and rules of netizens' online responses. The research shows the issues of earthquake donation and Guo Meimei event are the public focus, the emotional polarities of interactions between netizens are basically the same, and the number of negative posts first is more than that of positive posts first in the interactions. Due to the diversity and dynamic of events, it is very significant to visually express the evolution process of hot events. This paper proposes the method of extracting the storylines of the Red Cross relevant posts by combining LDA topic model and Bert vector. As a result, we obtain the development of the hot events, and the storylines reveal the trust crisis of the Red Cross.