图书情报工作 ›› 2013, Vol. 57 ›› Issue (22): 132-137.DOI: 10.7536/j.issn.0252-3116.2013.22.021

• 知识组织 • 上一篇    下一篇

高校图书馆馆藏资源协同推荐系统研究

邱均平, 张聪   

  1. 武汉大学信息管理学院
  • 收稿日期:2013-10-22 修回日期:2013-11-08 出版日期:2013-11-20 发布日期:2013-11-20
  • 通讯作者: 张聪,武汉大学信息管理学院硕士研究生,E-mail:agezhc@whu.edu.cn。
  • 作者简介:邱均平,武汉大学信息管理学院教授,博士生导师
  • 基金资助:

     本文系国家社会科学基金重大项目“基于语义的馆藏资源深度聚合与可视化展示研究”(项目编号:11&ZD152)研究成果之一。

Research on the Collaborative Recommendation System of University Library Resources

Qiu Junping, Zhang Cong   

  1. School of Information Management, Wuhan University, Wuhan 430072
  • Received:2013-10-22 Revised:2013-11-08 Online:2013-11-20 Published:2013-11-20

摘要:

以武汉大学为实际研究环境,针对高校图书馆环境下协同推荐的特殊性,在缺乏用户对馆藏资源评分数据的情况下提出主要根据借阅时间计算用户的偏好值,并辅助使用用户评分对该值进行调整。将用户的信息需求分为长期信息需求和短期信息需求,分别计算用户对馆藏资源的短期偏好值和长期偏好值,运用Apache Mahout构造馆藏资源的协同推荐系统,针对用户的两种需求进行推荐,并且采用邮件调查的方式对推荐结果进行了检验。

关键词: 馆藏资源, 协同推荐, 信息需求, 用户偏好

Abstract:

This paper takes Wuhan University as the practical research environment. Arming at the particularity of collaborative recommendation in university libraries and limited user rating data on library resources, it calculates users' preference value mainly in accordance with their loaning time and adjusts the value with assist of users' rating data. Users' information needs are divided into two categories of long-term information needs and short-term information needs. Then two kinds of corresponding preference value are calculated separately to get more accurate recommendation. Finally, by virtue of Apache Mahout, a collaborative filtering recommendation system is constructed to recommend books to meet two kinds of information needs, and the results are tested by email survey.

Key words: library resource, collaborative recommendation, information need, User Preference

中图分类号: