摘要
可视化同被引分析步骤主要包括选择数据源、确定分析的元素、计算同被引频次、对数据进行标准化处理、对数据进行降维和图示、对图形进行分析解释。这些步骤中的关键技术在于对获得的高维引文关系数据进行降维和图示,目前应用较多的主要有聚类分析、多维尺度分析、因子分析、自组织映射神经技术、寻径网络技术、最小生成树法、三角测量、力矢量布局算法等。其中寻径网络技术以其较好的图示效果得到较多的应用。
The general steps of co-citation visualization include chosing a data resource, defining the unit of analysis, calculating similarities between uinits, standizing raw data, dimensionality reduction and mapping. The key techniques in co-citation visualization processes are dimensionality reduction and maping which can derive useful representations of high dimenational citation data. The dimensionality reduction and maping techniques which had been used currently are cluster analysis, mutidimentioanal scaling, factor a- nalysis, self-organizing feature map, And pathfinder network scaling, minimum spanning tree, triangulation, and force directed placement and so on. And pathfinder network scaling is more frequently used because of its preferable visualizing effect.
出处
《图书情报工作》
CSSCI
北大核心
2008年第11期22-25,131,共5页
Library and Information Service
关键词
同被引分析
可视化
技术
co-citation analysis visualization techniques
作者简介
李运景,女,1969年生,副研究员,博士,发表论文30余篇;
侯汉清,男,1943年生,教授,博士生导师,发表论文150余篇,出版专著11部;
薛春香,女,1981年生,讲师,博士,发表论文10余篇;
任银玲,女,1964年生,副研究员,发表论文20余篇。