基于RelFinder的图情学科关联数据语义关系发现实践

发布时间：2019-03-08 19:15

【摘要】：[目的 /意义]对LISTA数据库收录的图情学科学术文献、期刊、作者的题录数据进行关联数据发布研究,并利用可视化软件Rel Finder进行多角度的语义关系发现实验,以期揭示不同学术单元数据之间的隐含关系和一些潜在的规律。[方法 /过程]首先,选取图情学科为研究领域,抓取LISTA数据库666种期刊、5 075位核心作者以及1 073篇学术文献的题录数据并导入MYSQL数据库。接着,构建轻量级任务本体对数据进行规范化描述,利用开源软件D2RQ转换为RDF三元组,并结合语义仓储软件Virtuoso发布为关联数据。最后,使用RDF可视化软件Rel Finder进行图情学科学术单元之间的语义关系发现,对直接关系、一次间接关系、二次间接关系以及多次间接关系的发现过程进行总结。[结果 /结论]Rel Finder能较好地发现图情学科学术文献、期刊、作者之间隐含的深层次关系,对检索学术文献、揭示学术脉络和发现学术领域知识都有重要意义。
[Abstract]:[objective] to publish the related data of academic documents, periodicals and authors' titles collected in the LISTA database, and to use the visualization software Rel Finder to discover the semantic relations from many angles. In order to reveal the relationship between different academic unit data and some potential laws. [methods / process] first of all, the subject of graph and situation was selected as the research field, and the title data of 666 journals, 5 075 core authors and 1 073 academic documents in LISTA database were captured and imported into MYSQL database. Then, a lightweight task ontology is constructed to standardize the description of the data, and the open source software D2RQ is used to transform the data into RDF triplets, and the semantic storage software Virtuoso is used to publish the associated data. Finally, we use RDF visualization software Rel Finder to discover the semantic relationships between academic units of graphic and emotional disciplines, and summarize the discovery process of direct relations, one indirect relations, two indirect relations and many indirect relationships. [results / conclusion] Rel Finder can well discover the deep-seated relationship among the academic documents, journals and authors, which is of great significance for searching the academic documents, revealing the academic context and discovering the knowledge in the academic field.
【作者单位】：北京师范大学政府管理学院;
【基金】：2016年度国家社会科学基金项目“基于语义识别的引文分析理论、方法与应用研究”(项目编号:16BTQ073)研究成果之一
【分类号】：G250.7;TP311.13

【相似文献】