摘要
将微博中的兴趣关注根据现有的类别进行再次分类。以新浪微博达人为研究对象,提取他们关注的名人以及机构,并将这些名人、机构根据主页描述和标签进行归类。基于共链关系统计同时关注每两个类别之间的用户人数。最后将统计结果制成相关性矩阵,导入SPSS软件中进行k-means聚类,结果为具有相似性的兴趣可以聚为一组。最后根据聚类结果结合现实情况分析各类别之间的相似性与区别。挖掘用户关注兴趣的隐性信息,并对微博用户推荐兴趣提出建议。
Taking the Sina microblog users as the object of research, this paper extracts those celebrities users, classifies them according to their home page tags, and counts the number of users between every two kinds of interests based on co-link relationship. According to the result of k-means cluster analysis, this paper finds the similarity between each kind of interest as well as the comprehensive information of users＂ interests, and finally gives some advice to the present attention recommending system.
出处
《情报杂志》
CSSCI
北大核心
2013年第8期142-144,131,共4页
Journal of Intelligence
