基于转发关系和单词特征的微博话题识别模型  

A microblog topic recognition model based on forwarding relationship and word feature

作　　者：钱晨嗣 陈伟鹤[1] QIAN Chen-si;CHEN Wei-he(School of Computer Science and Communication Engineering,JiangSu University,Zhenjiang 212013,Jiangsu Province,China)

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013

出　　处：《信息技术》2018年第9期44-49,54,共7页Information Technology

基　　金：江苏省六大人才高峰项目(2014_WLW_012);江苏省重点研发计划(社会发展)项目(BE2016630);江苏省重点研发计划(社会发展)项目(BE2015617)

摘　　要：相对一般文本,微博中包含大量的转发关系,传统的文本中挖掘算法不能很好地建模。单词具有情感特征,且微博话题标签更够揭示本中的主要内容。针对以上两点提出基于转发关系和单词特征的主题模型。该模型首先将用户主题模型和转发关系结合,然后根据单词特征求话题权重以新话题列表。实验表明该模型可能取得较好的效果。Compared with traditional text,microblog contains a large number of forwarding relationship,so traditional text mining algorithms cannot model it well. Words contain emotional characteristics and microblog hashtag is more likely to reveal the main content of the text. For the above two points,this paper puts forward a user theme model based on forwarding relationship and word features. The model first combines the user theme model with the forwarding relationship,and then according to the character of the word weight to update the microblog topic list. The experimental results show that this model can achieve more accurate prediction.

关 键 词：微博话题 转发关系 单词特征 情感特征 微博标签 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]