多源媒体文本主题演变的可视分析-计算机辅助设计与图形学学报.PDFVIP

29 12 Vol.29 No.12 2017 12 Journal of Computer-Aided Design Computer Graphics Dec. 2017 多源媒体文本主题演变的可视分析 张 怡, 邵裕东, 张加万 (天津大学软件学院 天津 300350) (yizhang@) : 传统新闻媒体和新兴社交媒体对同一事件的观点有差异, 且互相影响. 为了有效地分析不同来源的新闻事 件评论主题的异同以及随着时间推移互相影响的转移趋势, 提出一套可视分析方法. 首先提出一个基于主题分析的 信息传递模型, 使用 LDA 模型提取文本主题, 通过计算主题的内容相关性和时间相关性得到主题间的传递关系; 然 后使用基于桑基图和时间线技术相结合的方法对主题传递模型进行可视化, 并提供主题层级视图、关键词词距视图 和原数据视图帮助用户理解主题; 最后研发了可视分析系统原型. 使用“韩国部署萨德事件”这一案例, 验证了系统 的可用性和有效性. : 社会信息科学可视化; 可视化工具设计; 文本可视化; 时序数据分析; 主题可视化 : TP391 DOI: 10.3724/SP.J.1089.2017.16634 Visual Analysis of Topic Transition among Different Sources of Text Corpora Zhang Yi, Shao Yudong, and Zhang Jiawan (School of Computer Software, Tianjin University, Tianjin 300350) Abstract: Traditional news media and social media may convey different perspectives and topics may affect each other in the same event. This paper presents a set of visual analysis methods to analyze the similarities and differences in comments of event on multi text sources and the topic transition pattern over time. First, an information transition model based on topic analysis is proposed. LDA model was used to extract topics and the transmission relationships between topics are analyzed by content correlation and temporal correla- tion. An approach which combine Sankey graph and timeline technique is applied to visualize the topic tran- sition model. The topic hierarchy view, word distance view and raw data view are used to help users com- prehensive understanding the topics. Finally, this paper provides a visual analysis system prototype. The event of “South Korea deployed THAAD” is provided to verify the usabilit