图书情报工作 ›› 2020, Vol. 64 ›› Issue (9): 95-103.DOI: 10.13266/j.issn.0252-3116.2020.09.011

• 情报研究 • 上一篇    下一篇

融合内容与关系的学术社交媒体上跨学科用户推荐模型研究

吴小兰1, 章成志2   

  1. 1 南京师范大学新闻与传播学院网络与新媒体系 南京 210046;
    2 南京理工大学经济管理学院信息管理系 南京 210094
  • 收稿日期:2019-10-09 修回日期:2020-02-01 出版日期:2020-05-05 发布日期:2020-05-05
  • 通讯作者: 章成志(ORCID:0000-0001-9522-2914),教授,博士,博士生导师,通讯作者,E-mail:zhangcz@njust.edu.cn
  • 作者简介:吴小兰(ORCID:0000-0003-1869-1738),副教授,博士。
  • 基金资助:
    本文系国家社会科学青年基金项目"社交媒体视域下的跨学科用户发现及其推荐研究"(项目编号:17CTQ047)研究成果之一。

Research on Interdisciplinary User Recommendation Model in Academic Social Media Combining Content and Relations

Wu Xiaolan1, Zhang Chengzhi2   

  1. 1 Department of Internet and New Media, School of Journalism and Communication, Nanjing Normal University, Nanjing 210046;
    2 Department of Information Management, School of Economics&Management, Nanjing University of Science and Technology, Nanjing 210094
  • Received:2019-10-09 Revised:2020-02-01 Online:2020-05-05 Published:2020-05-05

摘要: [目的/意义] 在学术社交媒体快速发展的今天,开展跨学科研究或者寻求跨学科合作时,很多科研合作起始于社交媒体上的相识或关注,因此开展社交媒体上跨学科用户推荐非常有意义。社交媒体上主要存在"媒体"(代表内容)、"社交"(代表关系)两大类数据,因此本文开展了融合内容与关系的社交媒体跨学科用户推荐。[方法/过程] 在基于向量空间模型的用户表示之后,本文借助用户内容信息计算用户领域专业度,根据关系数据测度用户跨学科距离,同时结合用户关系网络PageRank值给出推荐结果。[结果/结论] 以科学网为例,实现"图书情报""计算机""新闻与传媒""高等教育""生物学"这5个领域内的跨学科用户推荐,并经人工实验测试检验,表明推荐结果在一定程度上能满足推荐需求。

关键词: 跨学科用户, 推荐模型, 跨学科距离, 学术社交媒体

Abstract: [Purpose/significance] With the rapid development of academic social media, when users do interdisciplinary research or seek interdisciplinary cooperation, many scientific research cooperation starts from the acquaintance or attention in social media, so it is very meaningful to research on interdisciplinary user recommendation in academic social media. There are two main types of data in social media:media (represents content) and social (represents relationship). Therefore, this paper recommended interdisciplinary users integrating content and relations. [Method/process] After user modeling based on Vector Space Model, this paper calculated user specialization with user content information, measured user's interdisciplinary distance based on relational data, then gave recommendation results combined with PageRank value of user relationship network. [Result/conclusion] Taking the science blog as an example, an interdisciplinary user recommendation model in five fields of "Library and Information", "Computer", "News and Media", "Higher Education" and "Biology" been implemented, which has been tested by artificial experiments, and showed that the recommendation results can meet the recommendation requirements to some extent.

Key words: interdisciplinary users, recommendation model, interdisciplinary distance, academic social networking

中图分类号: