:::

詳目顯示

回上一頁
題名:意見領袖關係及主題參與傾向研究--基於微博熱點事件的耦合分析
書刊名:新聞與傳播研究
作者:王晗嘯于德山
出版日期:2018
卷期:2018(1)
頁次:51-65+127
主題關鍵詞:意見領袖耦合分析因子分析社會網絡分析可視化
原始連結:連回原系統網址new window
相關次數:
  • 被引用次數被引用次數:期刊(0) 博士論文(0) 專書(0) 專書論文(0)
  • 排除自我引用排除自我引用:0
  • 共同引用共同引用:0
  • 點閱點閱:2
該文根據2016年486起微博熱點事件參與情況對325名意見領袖進行關聯,構建意見領袖事件耦合矩陣,綜合運用耦合分析、因子分析、社會網絡分析,分析耦合關系,以可視化的形式揭示意見領袖的活躍程度、彼此間的關系遠近以及主題參與傾向。研究發現,微博生態泛娛樂化與生活化傾向明顯。目前微博中主要活躍著四種意見領袖:休閑娛樂類意見領袖、社會民生類意見領袖、競技體育類意見領袖和金融時政類意見領袖。比較這些意見領袖及其與不同主題之間的關系可以發現,休閑娛樂類和社會民生類的意見領袖存在大量重疊的地方,這兩個主題的相關程度相對較高;金融時政類和競技體育類的意見領袖雖然在主題參與上與其他主題互有交織,但更多的是關注本主題下的事件,與其他主題的相關性相對較低。該研究通過耦合實現了文本與用戶的連接,對用戶間的隱性關系進行挖掘,相關發現對于營銷號集體行為識別、意見領袖標簽的動態劃分、主題分級監測等方面有一定價值。
In this research,participations of 486 Weibo hot events in 2016 are associated with 325 opinion leaders to build an opinion leaders’event coupling matrix.Using coupling analysis,index analysis and social network analysis,the study analyzes the coupling relationship and visually reveals the activity of opinion leaders,the relationship between them and their subject participation tendency.Research findings suggest that Weibo’s ecosystem has an obvious preference for pan-entertainment and everyday life.Currently,four types of opinion leaders are active in the Weibo ecosystem:(1)opinion leaders in recreation and entertainment;(2)opinion leaders in society and everyday life;(3)opinion leaders in competitive sports;and(4)opinion leaders in finance and politics.Through comparing these opinion leaders as wel as relationships between them and different subjects,it was found that there are many overlaps between opinion leaders of types(1)and(2);the degree of correlation between these two subjects is relatively high;and in terms of opinion leaders in types(3)and(4),although their subjects may involve other subjects,the degree of correlation between their subjects and other subjects is relatively low,implying that they only focus on events in their own subject in most cases.This study uses coupling to realize the connection between the text and the user and explore the hidden relationship between the users.Its results may be used for collective behavior identification of marketing accounts,label dynamic division of opinion leaders and grading detection of subjects.
 
 
 
 
第一頁 上一頁 下一頁 最後一頁 top