图书情报知识 ›› 2017, Vol. 0 ›› Issue (2): 75-82.doi: 10.13366/j.dik.2017.02.075

• 情报、信息与共享 • 上一篇    下一篇

作者共引分析方法的扩展与效能改进研究

黄文彬,步一,王冰璐   

  • 出版日期:2017-03-10 发布日期:2017-03-10

Towards Improvements of Raw Co-Citation Matrix in Author Co-Citation Analysis

  • Online:2017-03-10 Published:2017-03-10

摘要:

本文从文献共引分析的研究视角改进了传统作者共引分析方法的作者频次计算方法,提出了“扩展ACA方法”,增加共引矩阵的信息量,构建共引矩阵与绘制分析该领域的科学图谱,借此改进传统方法上输入信息量过小、附加题录信息利用过少等缺陷。笔者主要是将引文发表时间信息和扩展作者共引关系依据不同权重计算带时间信息的共引矩阵,转化为相关矩阵后,通过GePhi软件绘制图谱并进行数据解释与分析。结果显示,综合时间信息的扩展ACA方法不仅保留了传统ACA的能力,而且提高了作者的聚类结果和知识图谱的可视化程度,挖掘出科学共同体的更多细节。

关键词: 作者共引分析, 共引分析, 引文分析, 信息计量学

Abstract:

We propose documentbased ACA (DACA) which utilizes the co-citation counts of the same pair of authors’ articles as the inputs of raw co-citation matrix to reconstruct it. Besides, we improve the method of calculating citations published time with the hypothesis that papers published in a short time are likely to be ignored because of its relatively low citations. This paper puts citations published time and documentbased co-citation into traditional ACA, namely documentand timebased ACA (DTACA). Network visualization and MDS-measurement are then employed. The results show that the proposed DTACA method can mine more details in authors’ clustering and can thus improve the performance and the accuracy of knowledge domain mapping, compared to traditional ACA.

Key words: Author Co-citation analysis, Co-citation analysis, Citation analysis, Informetrics