基于用户影响力的热点话题检测方法研究

发布时间：2018-02-04 04:43

  本文关键词： 话题挖掘 用户影响力 微博 文本挖掘　出处：《情报杂志》2017年04期 　论文类型：期刊论文

【摘要】：[目的/意义]对微博消息进行热点话题挖掘,进而从海量微博文本中实时找出用户关注、讨论的热点事件,是进行舆情监测、应急管理的基础。然而,现有微博热点话题检测研究却大多忽略了不同影响力用户对话题产生及传播的作用,并且检测结果直观性较差。针对此问题,提出了基于用户影响力的热点话题检测方法。[方法/过程]首先识别用户特征要素,构建用户影响力模型,计算用户影响力;然后,综合考虑主题词影响力、影响力增长速度和增长斜率,提出基于用户影响力的微博热点话题主题词抽取方法,抽取主题词簇;之后,识别核心主题词并进行热点话题关键词抽取。最后,通过实验验证方法的有效性。[结果/结论]实验结果表明:基于用户影响力的热点话题检测方法能够有效识别并直观表达出检测时间窗口内的典型热点话题;该方法能有效提升实证性热点话题识别效率,减少娱乐性热点话题的识别;通过对不同时间窗口内同一话题的关键词抽取,可以实现对相应话题的热点跟踪。
[Abstract]:[Objective / significance] to mine the hot topic of Weibo message, and then to find out the user's attention and the hot events discussed in real time from the massive Weibo text, which is the basis of public opinion monitoring and emergency management. The existing research on Weibo hot topic detection mostly ignores the influence of different users on the topic generation and dissemination, and the detection results are not intuitive. A hot topic detection method based on user influence is proposed. [Method / process: firstly, the user characteristic elements are identified, the user influence model is constructed, and the user influence is calculated. Then, considering the influence of theme words, the speed of influence growth and the slope of growth, a method of extracting the subject words of Weibo hot topic based on user influence is put forward to extract the cluster of theme words. After that, the key words are identified and the key words of hot topics are extracted. Finally, the effectiveness of the method is verified by experiments. [Results / conclusion] the experimental results show that the hot topic detection method based on user's influence can effectively identify and express the typical hot topic in the detection time window. The method can effectively improve the efficiency of the empirical hot topic identification and reduce the entertainment hot topic recognition. By extracting the keywords of the same topic in different time windows, the hot spot tracking can be realized.
【作者单位】： 大连理工大学管理与经济学部;
【基金】：辽宁省社会科学规划基金重点项目“突发事件网络舆情的动态监测与预警策略研究”(编号:L15AGL017) 国家自然科学基金项目“在线知识社区中社会系统与知识系统协同序化机制和规律研究”(编号:71573030)的研究成果之一
【分类号】：TP393.092;TP391.1
【正文快照】： 关键词话题挖掘用户影响力微博文本挖掘引用格式裘江南,谷文静,翟R，

本文编号：1489356