摘要
语言作为逻辑思维和推理工具,其基本要素是语词。重要语词映射成概念,核心概念建构知识,而知识演进促成学术发展。本文以学术文献中的关键词作为概念演化基础,尝试超越传统引文分析法探讨测量学术贡献的关键词分析法。研究发现以一个学科内的关键词向量和关键词流量为基准,可分别定义关键词向量通量谱和累积通量谱的h截断和g截断,构成h核和g核内主流关键词,进而可定义主流率和主流指数,用作测量对主流学术研究之学术贡献的新型学术测评参数。合并CNKI、万方、维普三大中文数据库元数据和WoS、Ei里的中文记录元数据建立了2008-2017年数据支持平台,具体探讨了针对学术团体和学术期刊的中观层面实证案例。本文给出的关键词分析法基础框架重点是挖掘学术主体或客体对主流研究的学术贡献,建议在试用和应用中进一步改良和完善。
As a tool of logical thinking and derivation,the basic element of language is the word. Important words reflect concepts,and core concepts construct knowledge,while knowledge evolution contributes to academic development. Based on the theoretical foundation that keywords in the academic literature characterize concepts,this work attempts to get beyond traditional citation analysis and introduce a framework of academic contribution analysis based on keywords.The framework of keyword analytics defines keyword vector flux spectrum and cumulative flux spectrum based on the keyword vector and keyword flow in a discipline,and then applies h-index and g-index to measure h-cutoff and g-cutoff,which constitute the mainstream keyword set in h-core and g-core. We then define the mainstream ratio and mainstream index,which can be used to measure the contribution of academic subjects( e.g. institutions) or academic objects( e.g. journals) to mainstream academic research.The method of keyword analytics for measuring academic contributions inherits the theoretical characteristics of h-type metrics,and provides a new measurement in addition to citation analytics. In the field of humanities,the traditional quantitative evaluation methods have many limitations,and the keyword analysis method is worthy to explore. This method framework still has limitations and needs further innovation and development. For example,a series of conceptual clues such as keyword,theme-word,concept-word and ontology,are currently confused and need to be clarified. While using keyword analytics,the techniques of front-end word processing and back-end text mining still need to be further explored. On the basis of semantics,how to measure synonyms,near-synonyms,substitute words,and evolution of keyword itself accurately,is also a problem and needs further exploration.The empirical data in this article is derived from the data platform built by the authors. The platform consists of three Chinese databases,CNKI,Wanfang and Weipu,and Chinese records in Wo S and Ei. The data is cleaned and processed by machine plus manual work. The platform includes 87 million Chinese achievement data and 17 million foreign language achievement data,totaling 104 million,with a time span of 2008 to 2017. In this study,two data subsets on library and information science and philosophy were respectively selected as empirical cases.
作者
黄晨
赵星
卞杨奕
张家榕
张慧
叶鹰
C.HUANG;X.ZHAO;BIAN Yangyi;Ronda J.ZHANG;H.ZHANG;Y.YE
出处
《中国图书馆学报》
CSSCI
北大核心
2019年第6期84-99,共16页
Journal of Library Science in China
