计算机与现代化

• 网络与通信 • 上一篇    下一篇

基于用户关联的热点话题检测方法

  

  1. 南京航空航天大学计算机科学与技术学院,江苏南京210016
  • 收稿日期:2014-12-02 出版日期:2015-04-27 发布日期:2015-04-29
  • 作者简介:李洪利(1990-),女,江苏滨海人,南京航空航天大学计算机科学与技术学院硕士研究生,研究方向:信息安全,信息检索; 王箭(1968-),男,教授,博士生导师,研究方向:应用密码学,系统安全分析与设计 。

Hot Topics Detection Method Based on User Relevance

  1. College of Computer Science & Technology, Nanjing University of Aeronautics & Astronautics, Nanjing 210016, China
  • Received:2014-12-02 Online:2015-04-27 Published:2015-04-29

摘要:

为了提高从微博信息中检测热点话题的准确率,提出一种基于用户关联的热点话题检测方法。依据用户权威度对微博用户排序,过滤拥有低权威度值用户的相关微博信息。用户权威度通过
基于PageRank的微博用户权威度评价模型来计算。将用户权威度数据与微博评论数、转发数结合在一起,得到话题的热度值。实验结果表明,所提出的方法在微博应用中能够有效识别“僵尸”用户,使
漏检率和误检率分别平均降低19.78%和4.77%,能较好地提高热点话题检测成果的准确率。

关键词:  , PageRank算法, 微博, 用户权威度, 话题检测

Abstract:

To improve the accuracy of detecting hot topic from social network text information, a hot topic detection algorithm based on user relevance is raised.
According to user authority to the user orientation, some noise data of parts of users are filtered out. The authority of users is calculated by micro-blog users authority
evaluation model based on PageRank algorithm. The heat evaluation of micro-blog topics is made by combining the number of reposts and comments with user authority data together
, thus the hot topics are found. The result of the experiment shows that the method can effectively identify “zombie” users, and further the average missing rate and false
detection rate respectively decrease by 19.78% and 4.77%, indicate the topic detection accuracy rate is effectively improved.

Key words: PageRank algorithm, micro-blog, user authority, topic detection