图书情报工作 ›› 2019, Vol. 63 ›› Issue (9): 61-72.DOI: 10.13266/j.issn.0252-3116.2019.09.007

• 情报研究 • 上一篇    下一篇

基于科技文献多重共现的数据模型理论与知识发现应用范例研究

庞弘燊   

  1. 深圳大学图书馆 深圳 518060
  • 收稿日期:2018-02-27 修回日期:2018-07-23 出版日期:2019-05-05 发布日期:2019-05-05
  • 作者简介:庞弘燊(ORCID:0000-0002-5039-8817),副研究馆员,博士,E-mail:phs@szu.edu.cn。
  • 基金资助:
    本文系教育部人文社会科学研究青年基金项目"基于科技文献多源数据融合及多特征项共现分析技术的科研动态识别监测方法研究"(项目编号:18YJC870015)和广东省哲学社会科学一般项目"学科领域创新演化路径的情报分析方法研究"(项目编号:GD18CTS03)研究成果之一。

Research on Data Model Theory and Knowledge Discovery Application Based on Multiple Occurrence of Scientific Literature

Pang Hongshen   

  1. Library, Shenzhen University, Shenzhen 518060
  • Received:2018-02-27 Revised:2018-07-23 Online:2019-05-05 Published:2019-05-05

摘要: [目的/意义]科技文献中各种特征项及其之间的关联是构成多种多样共现现象的基本单元,通过挖掘共现特征项之间的关联,共现分析可以从不同角度探测科学与技术活动规律的方方面面,为科研管理者和研究者等提供一个全方位、多角度观察科学发展的新视角。[方法/过程]通过对多重共现的基础理论研究,构建一套独特的多重共现数据模型基础理论体系,该理论体系包括:多重共现的定义、多重共现的研究范畴、用于多重共现的变量符号、多重共现的矩阵定义、多重共现的数据组织形式以及多重共现的延展系数计算公式与应用范畴。此外,基于多重共现的交叉图可视化方式,构建可用于分析3个或以上特征项共现关系的知识发现方法,包括共现关联强度、被引关联强度以及共现突发强度的分析方法。[结果/结论]通过该基础理论体系的构建,拓展共现现象的研究范围,为共现分析走向多角度、多维度的多重共现分析提供基础理论的支持。并通过实证研究,选取不同的多重共现应用案例,证明该方法可应用于研究领域、研究机构、机构间对比、研究学者等方面的分析,同时具有较好的分析效果。由于该方法体系具有分析角度多维化和分析方法多样化的特点,通过该方法的分析,除能够实现一重、二重共现等的分析效果外,还能揭示出比一般共现更为广泛和深入的知识内容。

关键词: 多重共现, 多特征项共现, 多源数据, 数据模型, 知识发现

Abstract: [Purpose/significance] Various entities and their associations are the basic units that constitute a variety of occurrence phenomena in scientific literature. By mining the associations between occurrence entities, occurrence analysis can detect all aspects of the laws of scientific activities from different angles for scientific research management and researchers. It will provide a new perspective on the development of science from all angles and perspectives.[Method/process] By studying the basic theory of multiple occurrence, this paper constructs a set of unique basic theoretical system of multiple occurrence data model. The theoretical system includes definition of multiple occurrence, multiple occurrence research category, multiple occurrence variable symbols, multiple occurrence matrix definitions, multiple occurrence data organization forms, etc. In addition, based on the multiple occurrence cross-graph visualization method, this paper constructs a knowledge discovery method that can be used to analyze the occurrence relationship of three or more characteristic items, including the occurrence relevance strength, cited relevance strength and occurrence burst strength method.[Result/conclusion] Through the construction of this basic theoretical system, the research scope of occurrence phenomena is expanded, which provides the basic theory support for occurrence analysis to multi-angle and multi-dimension occurrence analysis. And through empirical research, different cases of multiple occurrence applications are selected, proving that the method can be applied to the analysis of research areas, research institutions, institutional contrast, research scholars, etc., and has good analysis results. Due to the multi-dimensional analysis and the diversification of analysis methods, this method can not only achieve the analysis effects of occurrence which includes one entity or two entities, but also reveal more extensive than the common occurrence and in-depth knowledge of content.

Key words: multiple occurrence, multiple feature items occurrence, multi-source data, data model, knowledge discovery

中图分类号: