图书情报工作 ›› 2015, Vol. 59 ›› Issue (13): 126-133,148.DOI: 10.13266/j.issn.0252-3116.2015.13.018

• 知识组织 • 上一篇    下一篇

惩罚性矩阵分解及其在共词分析中的应用

邵作运1, 李秀霞2   

  1. 1. 曲阜师范大学日照校区图书馆 日照 276826;
    2. 曲阜师范大学传媒学院 日照 276826
  • 收稿日期:2015-05-18 修回日期:2015-06-05 出版日期:2015-07-05 发布日期:2015-07-05
  • 通讯作者: 李秀霞(ORCID:0000-0002-3492-4768),副教授,硕士生导师,E-mail:zyshao@126.com
  • 作者简介:邵作运(ORCID:0000-0003-1818-5587),馆员,硕士

Penalized Matrix Decomposition and Its Application in Co-word Analysis

Shao Zuoyun1, Li Xiuxia2   

  1. 1. Library of Rizhao Campus, Qufu Normal University, Rizhao 276826;
    2. School of Communication, Qufu Normal University, Rizhao 276826
  • Received:2015-05-18 Revised:2015-06-05 Online:2015-07-05 Published:2015-07-05

摘要:

[目的/意义] 基于高维矩阵稀疏降维的思想,提出一种利用惩罚性矩阵分解(Penalized Matrix Decomposition,PMD)实现共词分析的新方法。[方法/过程] 以"学科服务"为研究主题,根据PMD算法原理,在Matlab环境下分别实现特征词的提取、特征词的软聚类以及聚类效果的可视化。[结果/结论] 与传统的共词分析方法对比,PMD算法在共词分析中具有独特的优势:提取的特征词比较全面,聚类数目便于确定,聚类结果易于理解。

关键词: PMD算法, 共词分析, 特征词提取, 特征词软聚类, 可视化

Abstract:

[Purpose/significance] Based on the idea of sparse dimension reduction, this paper proposes a new co-word analysis method with PMD (Penalized Matrix Decomposition).[Method/process] According to the PMD algorithm principle, this paper takes the subject service as research theme, and separately extracts the feature words extracting, makes the feature words soft clustering and visualizes clustering results in the Matlab environment. [Result/conclusion] Comparing with the traditional co-word analysis method, this paper finds that the PMD algorithm has some unique advantages in the co-word analysis, it can extract characteristic words more comprehensively, easily determine the clustering number, and get the more well clustering results.

Key words: PMD algorithm, co-word analysis, feature words extraction, feature words soft clustering, visualization

中图分类号: