图书情报工作 ›› 2019, Vol. 63 ›› Issue (7): 96-104.DOI: 10.13266/j.issn.0252-3116.2019.07.012

• 知识组织 • 上一篇    下一篇

基于特征分解的知识网络结构关系提取

栾宇, 滕广青, 安宁, 韩尚轩   

  1. 东北师范大学信息科学与技术学院 长春 130117
  • 收稿日期:2018-07-10 修回日期:2018-11-08 出版日期:2019-04-05 发布日期:2019-04-05
  • 通讯作者: 滕广青(ORCID:0000-0002-1053-0959),教授,博士生导师,通讯作者,E-mail:tengguangqing@163.com
  • 作者简介:栾宇(ORCID:0000-0003-3752-2470),硕士研究生;安宁(ORCID:0000-0002-9579-0150),硕士研究生;韩尚轩(ORCID:0000-0001-0962-3218),博士研究生。
  • 基金资助:
    本文系国家自然科学基金面上项目"基于网络结构演化的Folksonomy模式中社群知识组织与知识涌现研究"(项目编号:71473035)研究成果之一。

Structural Relationships Extraction of Knowledge Networks Based on Eigen Decomposition

Luan Yu, Teng Guangqing, An Ning, Han Shangxuan   

  1. School of Information Science and Technology, Northeast Normal University, Changchun 130117
  • Received:2018-07-10 Revised:2018-11-08 Online:2019-04-05 Published:2019-04-05

摘要: [目的/意义]对知识网络中结构关系的有效识别与提取,有助于从纷繁的数据中探测知识网络的拓扑结构及其演化模式。[方法/过程]本文提出一种基于邻接矩阵特征分解的知识网络结构关系提取方法。基于真实数据分别从静态结构关系提取和动态结构演化两个方面,对特征分解法和传统关联频度法进行对比分析,并与Pathfinder算法进行对比。对基于特征分解法提取知识网络结构关系的有效性进行验证。[结果/结论]研究结果表明:特征分解法能够识别原始知识网络中的主要成分信息,能够准确识别低频次的对网络整体拓扑结构较为重要的关联关系,且提取方法灵活自由。

关键词: 知识网络, 特征网络, 特征分解, 结构关系

Abstract: [Purpose/significance] The effective identification and extraction of structural relationships in knowledge networks helps to detect the topology of knowledge networks and their evolution patterns from a wide range of data. [Method/process] This article proposes a method for extracting structural relationships in knowledge networks based on eigen decomposition of adjacency matrix. Using the real data, the eigen decomposition method and traditional correlation frequency method are compared and analyzed from static structural relationships extraction and dynamic structure evolution, and compared with the pathfinder algorithm. The validity of structural relationships extraction of knowledge networks based on eigen decomposition method is verified. [Result/conclusion] The research results show:the eigen decomposition method can identify the main component information in the original knowledge networks, the method can accurately identify the low-frequency correlations that are important to the global topology of the networks, and the extraction method is flexible and free.

Key words: knowledge network, eigen network, eigen decomposition, structural relationship

中图分类号: