图书情报工作 ›› 2019, Vol. 63 ›› Issue (23): 122-130.DOI: 10.13266/j.issn.0252-3116.2019.23.014

• 综述述评 • 上一篇    下一篇

社区画像研究综述

刘蕾蕾1,2, 王胜涛3, 胡正银1,2   

  1. 1. 中国科学院成都文献情报中心 成都 610041;
    2. 中国科学院大学经济与管理学院图书情报与档案管理系 北京 100190;
    3. 江南大学糖化学与生物技术教育部重点实验室 无锡 214122
  • 收稿日期:2019-04-01 修回日期:2019-06-18 出版日期:2019-12-05 发布日期:2019-12-05
  • 通讯作者: 胡正银(ORCID:0000-0002-5699-9891),副研究员,通讯作者,E-mail:huzy@clas.ac.cn
  • 作者简介:刘蕾蕾(ORCID:0000-0002-7269-5855),硕士研究生;王胜涛(ORCID:0000-0001-7883-3924),博士研究生。
  • 基金资助:
    本文系中国科学院"十三五"信息化专项"面向干细胞领域知识发现的科研信息化应用"(项目编号:XXH13506-203)研究成果之一。

A Literature Review on Community Profiling

Liu Leilei1,2, Wang Shengtao3, Hu Zhengyin1,2   

  1. 1. Chengdu Documentation and Information Center, Chinese Academy of Sciences, Chengdu 610041;
    2. Department of Library, Information and Archives Management, School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190;
    3. Key Laboratory of Carbohydrate Chemistry&Biotechnology, Ministry of Education, Jiangnan University, Wuxi 214122
  • Received:2019-04-01 Revised:2019-06-18 Online:2019-12-05 Published:2019-12-05

摘要: [目的/意义] 社区画像对于解决社交网络信息过载问题,实现深层次的个性化知识服务意义重大。针对社区画像研究现状,进行客观的分析与评价,以期为社区画像进一步研究与应用提供思路。[方法/过程] 通过文献调研与分析,从研究内容、方法体系和应用场景3方面对社区画像进行调研、分析与归纳,评述其研究现状,提出未来的重点研究方向。[结果/结论] 以分析静态用户数据,采用相似性方法画像为主,聚焦于推荐服务、社区发现等传统应用。当前社区画像研究尚处在起步阶段,其数据对象、研究方法与技术手段都有待丰富,社区画像的发展前景与应用空间广阔,需进一步开拓。

关键词: 社区画像, 用户数据, 内容画像, 传播画像, 社区发现, 推荐系统, 知识服务

Abstract: [Purpose/significance] Community profiling is important for solving the overload of social network information and helping to achieve personalized and deep knowledge services. This literature review presents the research status in community profiling, and analyzes the corresponding techniques, methods and applications, and aims to provide ideas for further research and application of community profiling.[Method/process] Based on the literature investigation, this paper reviews community profiling from three aspects:research content, techniques and methods, and application scenarios. Moreover, the key features and weaknesses of the discussed techniques and methods are presented and several key research fields for future research are highlighted.[Result/conclusion] It is found that the present research focuses on static user data, user similarity methods for profiling, and traditional applications such as recommended services and community discovery. At present, the research on community profiling is still in its infancy, and the data, techniques and methods need to be enriched. It should have good prospects and wide application in the future.

Key words: community profiling, user data, content profile, diffusion profile, community detection, recommender system, knowledge service

中图分类号: