摘要 [目的/意义]提取突发事件网络舆情中事件的因果和顺承关系,构建事理图谱揭示事理发展脉络和演化规律。[方法/过程]以微博相关评论为研究对象,分别通过因果关系标识、事件影响因子和时间因子等维度识别事件间的关系,提取因果和顺承事件对;采用层次聚类将相似度较高的事件泛化为一类,构建新冠肺炎网络舆情事理图谱,并对生成的舆情演化路径进行分析。[结果/结论]研究表明,因果关系演化路径具有事件少、演化路径短等规律性特点,且同时包含顺承关系事件和伪结果事件;顺承关系演化路径具有事件数量多、演化路径较长、交叠变化等规律性特点,且能够揭示舆情演化的关键节点。通过构建事理图谱可以快速区分突发事件舆情的噪音和关键信息,充分体现其在监测、预测和管理中的价值性。 [Purpose/significance]Extract the causal and sequential relationship of events in the network public opinion of emergencies,construct the event evolutionary graph,and reveal the development context and evolution law of the event.[Method/process]Taking Weibo related comments as the research object,the relationship between events is identified through dimensions such as causal relationship identification,event impact factor and time factor,and causal and continuity event pairs are extracted.Hierarchical clustering is used to generalize the events with higher similarity into one category,construct the COVID-19 network public opinion event evolutionary graph,and analyze the generated public opinion evolution path.[Result/conclusion]The results show that the evolution path of causality has the characteristics of few events and short evolution path,and it also includes the events of consequent relationship and pseudo consequence;the evolution path of consequent relationship has the regular characteristics of large number of events,long evolution path and overlapping change,and can reveal the key nodes of public opinion evolution.Through the construction of the event evolutionary graph,can quickly distinguish the public opinion noise and key information of emergencies,and fully reflect its value in monitoring,forecasting and management.
机构地区 广西民族师范学院经济与管理学院 天津大学教育学院
出处 《情报理论与实践》 北大核心 2021年第3期76-83,共8页 Information Studies:Theory & Application
基金 国家社会科学基金重大项目“大数据驱动的社交网络舆情主题图谱构建及调控策略研究”的成果,项目编号:18ZDA310。
关键词 新型冠状病毒肺炎疫情 事理图谱 网络舆情 演化路径 突发事件 舆情管理 COVID-19 event evolutionary graph network public opinion evolution path emergency events public opinion management
分类号 G63 [文化科学—教育学]
