[目的/意义]文献引用中普遍存在漏引现象，分析文献漏引行为的表现特征，有助于揭示不合理引用行为背后的引用动机与情境，避免漏引对学术论文价值造成的不良影响。[方法/过程]构建文献漏引概念模型，以Web of Science核心合集数据库中自然科学、社会科学和人文科学共15个学科的抽样文献作为数据来源，通过文本相似度与引用耦合强度确定可供分析的数据，利用非参数检验的方法从文献题录、文献内容、作者和期刊四个方面探索漏引行为表现特征。[结果/结论]结果显示，漏引行为在语言、学科、文献类型、被引频次、作者自引、作者H指数、期刊影响因子上表现特征明显，在高被引作者上表现特征则视领域而定，在文本内容相似度上表现特征并不明显，并推测表现明显的特征可能一定程度上影响了潜在漏引行为。[创新/局限]本文定量捕捉漏引行为的表现特征，有助于探索定量化研究文献漏引的工作，为基于大规模文献数据的文献漏引识别提供了参考。
Analysis of the Performance Characteristics of Literature Missing Citations Behavior Based on Non-Parametric Test
[Purpose/Significance]The phenomenon of missing citations is common in document citation.Analyzing the characteristics of missing citations can help reveal the motivation and context behind unreasonable citation behavior,and avoid the negative impact of missing citations on the value of academic papers.[Method/Process]Construct a conceptual model of literature missing citation,use the sample literature from 15 disciplines in the Web of Science core collection database,including nature science,social science,and human science as the data source,determine the data available for analysis through text similarity and citation coupling strength,and explore the performance characteristics of omission behavior from four aspects:literature titles,literature content,authors,and journals using non-parametric testing.[Results/Conclusion]The results show that missing citations behavior has significant characteristics in language,discipline,document type,citation frequency,author self citation,author H index and journal impact factors.For highly cited authors,the performance characteristics depend on the domain.It is not obvious in text content similarity.It is suspected that ob-vious characteristics may affect potential missing citations behavior to some extent.[Innovation/Limitation]This paper quantitatively captures the performance characteristics of missing citations behavior,which is helpful to explore the work of quantitative research lit-erature missing citations,and provides a reference for literature missing citations identification based on large-scale literature data.
Literature Missing CitationsCitation Coupling StrengthText SimilarityNon-parametric TestCitation Feature
NETL
