科技论文引文的相关性、因果性与引文评价

Relevance, Causality and Evaluation of Citations in Scientific Papers

引文评价具备可量化、数据易得和衍生数据多样等优点,已在学术评价中得到了广泛应用。但近年来,一些知名学者、科学组织和科技出版机构纷纷质疑引文评价的合理性,使其应用价值受到巨大挑战。为了从引用行为对引用结果的影响和文献相关性差异2方面阐述传统引文数据用于学术评价的不合理性,借鉴大数据研究的思路,提出因果性引文的概念并分析其特征,从引文目的、功能、原则和传播效应4个方面论证因果性引文应用于学术评价的合理性,并以学术搜索平台Sematic Scholar为例,阐述人工智能可以依据具体引文功能不同对传统引文评价进行优化,从而有利于构建良好科技评价和编辑出版生态。

Citation evaluation has been widely used in academic evaluation due to its advantages such as quantifiable,easily available data and various derived data. However,in recent years,some well-known scholars,scientific organizations and scientific publishing institutions have questioned the rationality of citation evaluation,which makes its application value face great challenges. In order to expound the irrationality of traditional citation data used in academic evaluation from two aspects:the influence of citation behavior on citation results and the difference of literature relevance, referring to the thinking of big data research,this paper puts forward the concept of causal citation and analyzes its characteristics,and demonstrates the rationality of applying causal citation in academic evaluation from the four aspects of citation purpose,function,principle and communication effect. Taking Sematic Scholar as an example,it is expounded that artificial intelligence can optimize the traditional citation evaluation according to different citation functions,which is conducive to building a good science and technology evaluation and editing and publishing ecology.