图书情报工作 ›› 2015, Vol. 59 ›› Issue (4): 122-128.DOI: 10.13266/j.issn.0252-3116.2015.04.018

• 知识组织 • 上一篇    下一篇

关联数据驱动的查询扩展技术研究

田野1, 杨眉1, 祝忠明2, 张静蓓3   

  1. 1. 上海交通大学图书馆 上海 200240;
    2. 中国科学院兰州文献情报中心 兰州 730070;
    3. 上海外国语大学图书馆 上海 201620
  • 收稿日期:2014-11-03 修回日期:2015-01-15 出版日期:2015-02-20 发布日期:2015-02-20
  • 作者简介:田野(ORCID:0000-0001-5335-2673),助理馆员,硕士,E-mail:ytian@lib.sjtu.edu.cn;杨眉(ORCID:0000-0002-4282-6738),副研究员,博士;祝忠明(ORCID:0000-0002-2365-3050),研究员,博士生导师;张静蓓(ORCID:0000-0002-2439-5049),助理馆员,硕士。

Research of Linked Data-driven Query Expansion

Tian Ye1, Yang Mei1, Zhu Zhongming2, Zhang Jingbei3   

  1. 1. Shanghai Jiaotong University Library, Shanghai 200240;
    2. The Lanzhou Branch of National Science Library, Chinese Academy of Sciences, Lanzhou 730000;
    3. Shanghai International Studies University Library, Shanghai 201620
  • Received:2014-11-03 Revised:2015-01-15 Online:2015-02-20 Published:2015-02-20

摘要:

[目的/意义] 针对当前查询扩展技术面临的瓶颈,提出一种关联数据驱动的查询扩展方法,改善检索系统的查全率、查准率。[方法/过程] 将扩散激活理论应用到关联数据集中,使得在输入查询词搜索潜在语义实体时,对提取的查询词的语义特征在知识库中进行有特定机制的扩散和激活,最后对这些语义关联的候补概念进行收集,并利用推理机制进行筛选,得到更优的概念集。[结果/结论] 该方法能有效提高检索系统的查全率、查准率,证明了本文提出的技术的可行性、有效性。

关键词: 查询扩展, 关联数据, 激活扩散模型, DBpedia, WordNet

Abstract:

[Purpose/significance] The current query expansion faced technology bottleneck,this paper presented a linked data-driven query expansion to improve retrieval system's recall precision.[Method/process] Applied the spreading activation model to the linked data graph. When input query words and searched for potential semantic meaning of query terms, there was a specific feature extraction mechanism of diffusion and activation in the knowledge base. Finally the candidate concepts for these semantic association were collected.[Result/conclusion] This method can improve retrieval system's recall precision. The technical feasibility and effectiveness was demonstrated.

Key words: query expansion, linked data, spreading activation model, Dbpedia, WordNet

中图分类号: