面向Cytoscape平台的关联数据知识图谱概览抽取与可视化

发布时间：2018-03-17 23:13

本文选题：关联数据　切入点：知识图谱概览　出处：《数据分析与知识发现》2017年03期 　论文类型：期刊论文

【摘要】：【目的】为更方便地查询和利用各个领域的海量关联数据,提出一种关联数据知识图谱概览的生成方法,使得用户在查询前就能了解关联数据访问点的内部数据结构。【方法】通过SPARQL查询关联数据所包含的领域知识关系,针对每一个知识关系构建知识图谱概览三元组并形成初步的知识图谱概览,再抽取每个知识分类的知识图谱概览三元组并合并到前者形成完整的知识图谱概览。【结果】研发Cytoscape插件实现此方法,并进一步提供知识图谱概览可视化功能。【局限】不能处理匿名节点等复杂知识分类抽取。【结论】在生物医学领域分别进行单点抽取、关联"桥"和关联"包含"三项测试,测试结果表明该方法抽取速度快而稳定,抽取结果的查全率高,且不需要网络爬虫或额外的索引工作。
[Abstract]:[objective] to make it easier to query and utilize vast amounts of associated data in various fields, a method for generating an overview of the knowledge map of associated data is proposed. Enable the user to understand the internal data structure of the associated data access point before the query. [method] query the domain knowledge relationship contained in the associated data through SPARQL, For each knowledge relationship, an overview of the knowledge map is constructed and a preliminary overview of the knowledge map is formed. Then extract the knowledge map overview of each knowledge classification and merge it into the former to form a complete knowledge map overview. [results] the Cytoscape plug-in is developed to implement this method. Furthermore, the visualization function of knowledge map overview is provided. [limitations] can't deal with the classification extraction of complex knowledge such as anonymous nodes. [conclusion] in the biomedical field, single point extraction is carried out respectively, and the association "bridge" and "association" contain "three tests". The test results show that the proposed method is fast and stable, has high recall rate and does not require web crawler or additional indexing work.
【作者单位】：北京师范大学珠海分校管理学院;
【基金】：广东省高等学校优秀青年教师培养计划项目“面向大数据的生物通路本体知识图谱可视化研究”(项目编号:YQ2015239) 广东省自然科学基金项目“基于本体推理演化的财经大数据分析与预测研究”(项目编号:2016A030313386)的研究成果之一
【分类号】：TP391.3

【相似文献】