计算机工程与应用 ›› 2023, Vol. 59 ›› Issue (13): 33-48.DOI: 10.3778/j.issn.1002-8331.2209-0475

• 热点与综述 • 上一篇    下一篇

医学知识图谱构建技术及发展现状研究

黄贺瑄,王晓燕,顾正位,刘静,臧亚男,孙歆   

  1. 1.山东中医药大学 智能与信息工程学院,济南 250355
    2.山东中医药大学 药学院,济南 250355
  • 出版日期:2023-07-01 发布日期:2023-07-01

Research on Construction Technology and Development Status of Medical Knowledge Graph

HUANG Hexuan, WANG Xiaoyan, GU Zhengwei, LIU Jing, ZANG Yanan, SUN Xin   

  1. 1.College of Intelligence and Information Engineering, Shandong University of Traditional Chinese Medicine, Jinan 250355, China
    2.College of Pharmacy, Shandong University of Traditional Chinese Medicine, Jinan 250355, China
  • Online:2023-07-01 Published:2023-07-01

摘要: 知识图谱作为人工智能的重要分支,因其强大的语义处理能力和数据组织能力,可以全面整合医学概念、挖掘潜在医学知识,已成为医学智能化发展的重要手段。鉴于此,论述了医学知识图谱搭建中知识抽取、知识表示、知识融合、知识推理四个过程的最新方法及特点,深入研究并对比不同方法的优缺点,归纳各阶段常用数据集,梳理知识图谱在医学知识问答、临床辅助诊疗、中医知识挖掘及药物研究等方面的研究现状及各场景下的应用难点。最后总结现有医学知识图谱技术的局限性及面临的挑战,并对其未来发展进行展望。

关键词: 医学知识图谱, 深度学习, 知识抽取, 本体, 知识推理

Abstract: As an important branch of artificial intelligence, knowledge graph can realize comprehensive integration of medical concepts and mining potential medical knowledge due to its powerful semantic processing ability and data organization ability, which has become an important means for the development of medical intelligence. Based on this, the latest methods and features of the four processes of medical knowledge graph building:knowledge extraction, knowledge expression, knowledge fusion and knowledge reasoning are discussed, the advantages and disadvantages of different methods are deeply studied and compared, the commonly used datasets in each stage are summarized, the research status of knowledge graph in medical knowledge question and answer, clinical auxiliary diagnosis and treatment, knowledge mining of traditional Chinese medicine and drug research are  reviewed, the application difficulties in each scenario are analyzed. Finally, the limitations and challenges of the existing medical knowledge graph technology are summarized and its future development is prospected.

Key words: medical knowledge graph, deep learning, knowledge extraction, ontology, knowledge reasoning