词序视角下学术文本关键词分布特征及其差异研究

Characteristics and Differences of Keywords Distribution in Academic Texts from the Perspective of Keywords Ranking

[研究目的]为了更加有效地发挥关键词在科研成果记录传播、信息资源组织建设与知识检索服务等方面的作用,提出词序视角下学术文本关键词分布特征及其差异考察的指标体系及其量化方法。[研究方法]首先,从学术文本关键词的选词偏好特征、选词多样性特征和选词语义特征等3个维度,选取并构建5个关键词分布特征量化指标;然后,利用文献计量和统计分析的方法处理所采集的数据;最后,对数据进行描述统计分析与不同类型学术文本关键词分布特征对比分析。[研究结论]以人工智能研究主题为例,发现相关的学术文本关键词以4~5个为主,作者更倾向在词序靠前的位置选择来自题名的词组作为关键词。随着词序的增加,关键词的选词多样性随词序变化先增后减;作者会在靠前词序标注与该研究主题核心概念高度相关的主题词作为关键词选词来源;关键词词对之间呈现一定的语义相似度递减现象;不同类型学术文本在各个指标层面呈现一定的差异性。

[Research purpose]To make more effectively play the role of scientific research results record dissemination,information resource organization construction and knowledge retrieval services,we propose an index system and quantitative methods for investigating the distribution characteristics of keywords in academic texts from the perspective of keywords ranking.[Research method]Firstly,five quantitative indicators of keywords distribution characteristics are constructed from three dimensions:keywords selection preference characteristics,keywords diversity characteristics and keywords semantic characteristics.Then the collected data are processed by bibliometrics and statistical analysis.Finally,we perform descriptive statistical analysis and compare the distribution characteristics of keywords in different types of academic texts.[Research conclusion]The keywords of academic texts related to artificial intelligence research topics are mainly 4 to 5.The authors tend to choose the phrases from the title as keywords in the front of the keywords ranking.As keyword rankings increase,keywords diversity first increases and then decreases.The authors will mark the subject terms which are highly related to the core concepts of the research topic as academic text keywords in the front position.There is a decrease in the semantic similarity between keyword pairs.Different types of academic texts show certain differences at each index level.