信息技术与管理应用

当前社交网络舆情事件呈现复杂的群体结构与内容动态演化特征，传统分析方法因忽视群体互动和内在演化规律等局限性，难以有效揭示微博重大舆情传播规律与机制。本研究选取多条微博热点话题数据，运用人工智能知识图谱工具Gephi和语言智能工具LTP从群体结构和句法结构动态演化的双重视角，定量分析网络舆情演化规律。实验表明微博舆情传播用户群体结构呈现快速增加、短期维持、慢速消减的物理演化规律，该规律用于指导语言治理建议及规则设计。文本句法结构与用户群体结构密切相关。基于群体结构和句法结构联合演化分析，提出4条治理语言结构化规则以及2条治理建议，提出网络空间语言治理倡议，为网络舆情监测和治理提供了新的研究视角和方法指引，促进网络重大舆情高效治理。

At present, public opinion events on social media present complex group structures and dynamic evolution characteristics. Traditional analysis methods, due to their limitations such as ignoring group interactions and internal evolution laws, are difficult to effectively reveal the dissemination law mechanism of major public opinions on Weibo. This paper, based on data of Weibo hotspot topic, with artificial intelligence knowledge graph tool Gephi and language intelligence tool LTP made a quantitative and qualitative analysis of the evolution process of public opinion from the perspective of the dynamic evolution of group structure and syntactic structure. Results show that the dissemination of public opinion on Weibo has significant phrase characteristics. The group structure shows dynamic evolution of periodicity trend, and the changes in syntactic structure are closely related to user interaction behaviors. Then, based on dynamic structure evolution, it proposed four rules for government governance language suggestions and two suggestions, which would offer new analytical perspectives and methodological support for network public opinion governance to promote large-scale online public opinion security and high efficient governance.