智能引文数据平台的功能比较与趋势分析  

Functional Comparison and Trend Analysis of Smart Citation Data Platform

作　　者：宋丹辉 陶俊[2] SONG Danhui;TAO Jun

机构地区：[1]深圳大学图书馆 [2]西北大学公共管理学院

出　　处：《图书馆论坛》2023年第11期59-69,共11页Library Tribune

基　　金：国家社科青年项目“基于引文内容标注的引文数据开放关联模型及发布流程研究”(项目编号:17CTQ005)研究成果。

摘　　要：因应传统引文分析忽视引用情境信息,无法提炼引文的情感、意图、重要性等内容特征,不能为细粒度知识发现、语义检索、学术评价等提供支持,Semantic Scholar、Scite、COCI等智能引文数据平台在引文内容分析理论和人工智能技术支持下应运而生。文章选取55篇特定主题的开放获取论文作为实例,考察引文内容特征的揭示及应用,并与代表传统引文数据的Web of Science,以及代表未来发展方向的COCI进行对比,分析在基础算法、分析粒度、数据准确性、服务方式等方面的趋势,以期为选择合适引文内容分析工具、构建中文智能引文数据平台、开展细粒度知识发现和科研评价提供参考。In response to the fact that traditional citation analysis,which ignores the citation context information,fails to refine content features such as emotion,intention,and importance of citations,and cannot support the fine-grained knowledge discovery,semantic retrieval,and academic evaluation,etc.,Semantic Scholar,Scite,COCI and other intelligent citation data platforms have emerged with the support of citation content analysis theory and artificial intelligence technology.Taking 55 open access papers on specific topics for example,this paper examines the disclosure and application of citation content features.By comparing them with Web of Science,which represents traditional citation data,and COCI,which represents the future development of semantic citation,it analyzes the trends in basic algorithms,analysis granularity,data accuracy,and service methods,with a view to providing references for selecting appropriate citation content analysis tools,building Chinese intelligent citation data platforms,and conducting fine-grained knowledge discovery and scientific research evaluation.

关 键 词：引文内容特征 引文意图 引文情感 ChatGPT 

分 类 号：G353.1[文化科学—情报学]