首页 | 新闻公告 | 投稿须知 | 编委会 | 关于杂志 | 订阅 | 留言FAQ | 广告服务 | 相关链接 | 下载区 | 联系我们

基于STM和格兰杰因果分析的网络新闻媒体倾向研究
Research on media slant of online news based on structural topic model and Granger causality analysis
摘要点击 328  全文点击 0  投稿时间:2019-01-28  修订日期:2019-10-22
  查看/发表评论  下载PDF阅读器
中文关键词  结构化主题模型;格兰杰因果分析;共享单车新闻;媒体倾向
英文关键词  structural topic model(STM), Grange causality analysis, bike-sharing news, media slant
基金项目  国家自然科学基金项目(面上项目,重点项目,重大项目);研究阐释党的十九大精神国家社会科学基金专项;北京市哲学社会科学规划项目
作者单位邮编
吴俊 北京邮电大学经济管理学院 100876
欧阳书凡 北京邮电大学经济管理学院 
李晓华 中国社会科学院工业经济研究所 
中文摘要
      网络媒体报道倾向对互联网新闻舆情热点传播及公众认知有着重要影响。为揭示互联网媒体在新闻报道中存在语义和传播倾向,采用纳入文档协变量的结构化主题模型量化不同媒体报道热点主题的语义强度,同时结合格兰杰因果分析检验不同媒体对热点主题时间序列的影响。研究结果表明, 5家互联网媒体在报道共享单车热点新闻时,存在主题选择及传播倾向性。表现在与用户关系密切的话题,媒体间存在跟风造势的倾向,与用户关联不大的话题,媒体间则较少传播或单向传播。提出的融合结构化主题模型的格兰杰因果分析方法为测度网络舆情主体对网络舆情事件的演化影响提供了新手段。
英文摘要
      Subjectivity of internet media has an important influence on the spread of online news and public opinion. To investigate the potential media slant and its propagation characteristics during the news reporting, a covariate-based structural topic model is used to measure slant effect of different internet media on news topic prevalence. Granger causality analysis is further conducted to test the causality relationship between media-oriented topic time series. Data analysis results reveal that the slanted topic selection and news propagation do exist for the five major internet media when they report the bike-sharing news. Specifically, the more topics that are closely related to users there are, the more causal links and follow trends there exist among the media. In addition, no link or only one directional causal link is shown between the media when topics are not focused on users. This paper contributes to incorporate topic modeling and Granger causality test to measure media slant and its impact on hot news topics evolution.
关闭

版权所有 © 2007 《系统工程学报》
通讯地址:天津市卫津路92号天津大学25教学楼A区908室 邮编:300072
联系电话/传真:022-27403197 电子信箱: jse@tju.edu.cn