一种基于时间片的微博用户分类方法

发布时间：2018-03-21 00:38

本文选题：微博用户分类　切入点：时间片微元　出处：《小型微型计算机系统》2013年11期 　论文类型：期刊论文

【摘要】：针对微博用户分类问题提出时间片微元的概念,建立了时间片微元模型;对每个时间片内的微博所涉及到的用户进行研究得到时间片微元内部的用户兴趣度向量,最终整合所有时间片内的用户兴趣度向量,再对整个时间段内用户的兴趣度向量进行两次朴素贝叶斯分类,得到整个时间段内的用户分类.同时,对微博内容研究上规避了传统的单一的依靠系统标签形成用户网络的方式,结合了自然语言处理技术,提取用户兴趣方向,形成用户兴趣向量,然后对用户兴趣向量分析,采用改进的朴素贝叶斯分类算法进行用户分类.最后是对所提出的方法按详细步骤进行实验,研究结果表明,基于时间片的微博用户分类方法能有效对大规模的微博语料中所涉及到的用户进行较准确分类,为研究微博用户分类问题具有一定推动作用.
[Abstract]:According to Weibo user classification problem, the concept of time slice microelement is put forward, and the time slice microelement model is established, and the user interest vector inside the time slice microelement is obtained by studying the user involved in each time slice. Finally, the user interest vector of all time slices is integrated, and then the user interest vector of the whole time period is classified twice by naive Bayes, and the user classification in the whole time period is obtained. At the same time, Weibo content research evades the traditional single way of forming user network based on system tags, combines natural language processing technology, extracts user interest direction, forms user interest vector, and then analyzes user interest vector. The improved naive Bayesian classification algorithm is used to classify users. Finally, the experimental results show that the proposed method is based on detailed steps. Weibo user classification method based on time slice can effectively classify the users involved in the large-scale Weibo corpus, and it has a certain role in promoting the research on the user classification problem of Weibo.
【作者单位】：武汉大学计算机学院;徐州工业职业技术学院信息管理技术学院;武汉大学软件工程国家重点实验室;
【基金】：国家自然科学基金项目(61070083)资助
【分类号】：TP393.092

【参考文献】