图书情报工作 ›› 2015, Vol. 59 ›› Issue (13): 104-110.DOI: 10.13266/j.issn.0252-3116.2015.13.015

• 情报研究 • 上一篇    下一篇

基于LDA和社会网络中心度的研究生个性化检索推荐模型研究

叶春蕾, 邢燕丽   

  1. 北京农学院图书馆 北京 102206
  • 收稿日期:2015-05-17 修回日期:2015-06-20 出版日期:2015-07-05 发布日期:2015-07-05
  • 作者简介:叶春蕾(ORCID:0000-0002-8776-194X),副教授,博士研究生,E-mail:ycl19750318@126.com;邢燕丽,信息咨询部主任,研究馆员。
  • 基金资助:

    本文系北京农学院2015年研究生改革与发展项目"面向研究生教育的主题个性化服务研究与实践"(项目编号:5076516002/043)研究成果之一。

The Model of Personalized Information Retrieval Recommendation Based on LDA and Social Network Centrality Analysis

Ye Chunlei, Xing Yanli   

  1. Library of Beijing University of Agriculture, Beijing 102206
  • Received:2015-05-17 Revised:2015-06-20 Online:2015-07-05 Published:2015-07-05

摘要:

[目的/意义] 为了解决研究生用户面临的检索问题,提出一种基于LDA和社会网络中心度分析的个性化检索推荐模型。[方法/过程] 首先,该模型以研究生学科专业为个性化特征,并据此选择相应的数据源。其次,该模型使用LDA识别主题内容,以完成全面知识的展示。再次,该模型根据用户提交的检索词在相应的关键词-主题共现网络中进行社会网络中心度分析,以完成用户相关知识的推荐。[结果/结论] 实验表明,该模型能够很好地解决研究生检索中个性化特征、全面知识展示以及相关知识推荐三大问题,其有效性得到一定程度的验证。

关键词: LDA主题模型, 中心度分析, 共现网络, 个性化检索推荐

Abstract:

[Purpose/significance] In order to solve the retrieval problem of postgraduates, this paper proposes a personalized retrieval recommendation model based on LDA and the social network centrality analysis. [Method/process] Firstly, this model takes the discipline and specialty of postgraduate as the personalized characteristics, and selects the relevant literature data source according it. Secondly, this model uses the LDA to identify the topic content, and completes the display of the comprehensive knowledge. Thirdly, this model takes the social network centrality analysis in the keyword-topic co-occurrence network according to the retrieval word, and completes the relevant knowledge recommendation. [Result/conclusion] Experimental results show that this model can solve the three retrieval problems of postgraduates, such as the personalized features, the comprehensive knowledge display and the relevant knowledge recommendation. Its validity is verified to some extent.

Key words: LDA topic model, centrality analysis, co-occurrence network, personalized retrieval recommendation

中图分类号: