摘要
应用基于期刊"引文-文本"拟合相似度的混合聚类算法对Web of Science中SCI、SSCI、A&HCI收录的8 305余种期刊进行聚类分析,并将期刊混合聚类结果与基本科学指标ESI学科分类体系进行比较研究。对于聚类结果与ESI体系相一致的期刊,认为其具有较强的"学科关联度",定义为相应学科领域的"学科内核",其余部分则为"学科外围"。"内核期刊"界定了相应学科的核心知识范围,"外围期刊"则诠释了不同学科的分化和交叉融合等结构特征。
A hybrid clustering algorithm based on fitting similarity of ＂citation -text＂, is used to cluster 8 305 journals in SCI, SSCI and A＆HCI in Web of Science, to interpret the internal structure of science. Essential Science Indication （ESI） classification system is compared with the journal hybrid clustering structure. The ＂coincidence＂ of the two schemes is thought having strong ＂ subject correlation＂ and defined as the ＂ subject kernel＂ of corresponding disciplines, and the ＂ difference＂ is ＂ subject peripheral＂. The ＂ Core journals＂ define the core knowledge scope of corresponding, while the ＂ peripheral journals＂ interpret the structure characteristics such as differentiation and cross fusion of different subjects.
出处
《图书情报工作》
CSSCI
北大核心
2013年第3期78-84,共7页
Library and Information Service
基金
国家自然科学基金项目"基于混合聚类技术的交叉学科结构特征和演进趋势研究"(项目编号:71103064)研究成果之一
关键词
期刊
混合聚类
学科交叉
科学结构
journal hybrid clustering interdisciplinary structure of science
