计算机技术与发展2018，Vol.28Issue(5)：17-22,6.DOI:10.3969/j.issn.1673-629X.2018.05.005
基于文本属性的微博用户相似度研究
Research on Micro-blog User Similarity Based on Text Similarity
摘要
Abstract
Traditional similarity calculation method ignores the subjective information of the users,which is an important element that re-flects user's interest point.In order to fully describe the user's information,the user's background information and their interactive con-tent on the social platform should be considered.Therefore,we present a calculating method of Micro-blog user similarity,which is bound up with text similarity.The user similarity is mainly divided by the background similarity and interest similarity which is mainly deter-mined by the text similarity.The cosine similarity should be calculated after the word segmentation and TF-IDF.User similarity is also described by user's location,the device they use,the time they send Weibo,the text they re-post and the relationship between them.Fi-nally,the method uses AHP to determine the weight of each attribute and build an integrated similarity calculation model.Through the ex-periment,systematically compared with the calculating method of user similarity combined with text similarity and the one before impro-ving,the results show that the former increase the F1metric by 34.3%,which shows its superiority.关键词
微博/社交网络/用户相似度/文本相似度/余弦相似度/层次分析法Key words
Micro-blog/social network/user similarity/text similarity/cosine similarity/analytic hierarchy process分类
信息技术与安全科学引用本文复制引用
李梦洁,邵曦..基于文本属性的微博用户相似度研究[J].计算机技术与发展,2018,28(5):17-22,6.基金项目
国家自然科学青年基金项目(61401227) （61401227）
