张雨,吴俊.科技政策知识图谱构建研究[J].数字图书馆论坛,2021,(8):31~38 |
科技政策知识图谱构建研究 |
Research on the Construction of Science and Technology Policy Knowledge Graph |
投稿时间:2021-07-06 |
DOI:10.3772/j.issn.1673-2286.2021.08.005 |
中文关键词: 知识图谱;科技政策;本体;Bi-LSTM |
英文关键词: Knowledge Graph; Science and Technology Policy; Ontology; Bi-LSTM |
基金项目:本研究得到国家重点研发计划项目“基于模式创新的科技咨询服务平台研发与应用示范”(编号:2018YFB1403600)资助。 |
作者 | 单位 | 张雨 | 北京邮电大学经济管理学院 | 吴俊 | 北京邮电大学经济管理学院 |
|
摘要点击次数: 2792 |
全文下载次数: 5825 |
中文摘要: |
为助力广大中小企业快速查新,亟需使用人工智能手段对科技政策文本知识建模,构建基于知识图谱的结构化查询。本研究以采集到的全国26?660条科技政策文本为数据源,首先构建科技政策知识本体,之后通过Bi-LSTM深度学习模型完成三元组抽取,最后应用Neo4j图数据库完成知识存储与图谱化检索。所构建的科技政策知识图谱共有4万余个实体节点、15万余条关系,能够实现不同细粒度政策实体和关系的关联查询与可视化。这种基于科技政策本体构建政策知识图谱的方法既拓展了垂直领域知识图谱的新思路,也为开拓基于互联网的科技政策智能问答奠定基础。 |
英文摘要: |
In recent years, governments at all levels have introduced numerous policies on scientific and technological innovation. Faced with massive policies, it is difficult for enterprises to make full use of the data resources. In order to encourage enterprises to accurately locate and quickly search for policies, it is necessary to use artificial intelligence methods to achieve knowledge modeling and build structured queries based on domain knowledge graph. This paper takes 26 660 science and technology policies as data sources, constructs the domain ontology of science and technology policies, and completes knowledge extraction through Bi-LSTM deep learning model. Finally, the data are stored in the graph database Neo4j and the knowledge graph of science and technology policy is constructed. The constructed science and technology policy knowledge graph has more than 40 000 entity nodes and more than 150 000 relationships, which can realize the associated query and visual presentation of different fine-grained policy entities and relationships. The method proposed by the research enhances the new ideas of the knowledge graph of the vertical domain and lays the foundation for the intelligent question answering system for science and technology policies. |
查看全文
查看/发表评论 下载PDF阅读器 |
关闭 |