arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1402.1778cs.SIcs.CYphysics.data-anphysics.soc-ph
收藏

分析人类和文化对象的异质社会网络

Analysis of a heterogeneous social network of humans and cultural objects

Santa Agreste, Pasquale De Meo, Emilio Ferrara, Sebastiano Piccolo, Alessandro Provetti

摘要:

现代在线社交平台使其成员能够参与广泛的活动,如交友,加入群组,发布/评论资源等。在本文中,我们研究了用户可以参与的不同活动之间是否存在相关性。为了进行我们的分析,我们专注于aNobii,这是一个拥有全球图书读者用户群的社交平台,他们喜欢发布他们的阅读,给出评级,评论书籍并与朋友和其他读者讨论。aNobii呈现了一种异构结构:i)部分社交网络,具有用户对用户的交互,ii)部分兴趣网络,具有图书收藏的管理,以及iii)部分大众分类法,具有由用户标记的图书。我们分析了aNobii的一个完整的匿名快照,我们专注于用户可以执行的三个特定活动,即她的标记行为,她加入群组的倾向以及她编写愿望清单报告她计划阅读的书籍的能力。以这种方式,每个用户与基于标签、基于组和基于意愿列表的简档相关联。通过信息论工具(如熵和互信息)进行的实验分析表明,基于标签和基于组的配置文件通常比基于愿望列表的配置文件信息量更大。此外,我们发现,与同一用户相关联的三个配置文件之间的相关度往往是小的。因此,用户分析不能简化为只考虑任何一种类型的用户活动(尽管很重要),但关键是要纳入多个维度来有效地描述用户的偏好和行为。

英文摘要:

Modern online social platforms enable their members to be involved in a broad range of activities like getting friends, joining groups, posting/commenting resources and so on. In this paper we investigate whether a correlation emerges across the different activities a user can take part in. To perform our analysis we focused on aNobii, a social platform with a world-wide user base of book readers, who like to post their readings, give ratings, review books and discuss them with friends and fellow readers. aNobii presents a heterogeneous structure: i) part social network, with user-to-user interactions, ii) part interest network, with the management of book collections, and iii) part folksonomy, with books that are tagged by the users. We analyzed a complete and anonymized snapshot of aNobii and we focused on three specific activities a user can perform, namely her tagging behavior, her tendency to join groups and her aptitude to compile a wishlist reporting the books she is planning to read. In this way each user is associated with a tag-based, a group-based and a wishlist-based profile. Experimental analysis carried out by means of Information Theory tools like entropy and mutual information suggests that tag-based and group-based profiles are in general more informative than wishlist-based ones. Furthermore, we discover that the degree of correlation between the three profiles associated with the same user tend to be small. Hence, user profiling cannot be reduced to considering just any one type of user activity (although important) but it is crucial to incorporate multiple dimensions to effectively describe users preferences and behavior.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑