融合子图结构的医学知识推理方法综述  

Overview of Medical Knowledge Inference Methods with Fusion Subgraph Structures

作　　者：宋豪俊 李燕[1] 刘悦悦 何欣宇 SONG Haojun;LI Yan;LIU Yueyue;HE Xinyu(College of Medical Information Engineering,Gansu University of Traditional Chinese Medicine,Lanzhou 730101,China)

机构地区：[1]甘肃中医药大学医学信息工程学院,兰州730101

出　　处：《医学信息学杂志》2025年第1期63-68,92,共7页Journal of Medical Informatics

基　　金：中国高校产学研创新基金—蓝点分布式智能计算项目(项目编号:2021LDA09002)。

摘　　要：目的/意义综述融合子图结构的医学知识图谱推理方法,为后续相关研究提供参考。方法/过程阅读并分析相关文献,结合知识图谱及其推理相关知识,分析目前融合子图结构的知识推理代表模型的特点和局限性,对比其与各类知识推理方法在相关领域任务中的优势和不足,总结归纳此类推理方法在医学领域的应用现状和未来发展前景。结果/结论未来研究应致力于探索医学领域内不同模态信息的交互关系,以丰富推理信息,从而构建更加完善的医学知识图谱,为临床实际问题提供有效解决方案。Purpose/Significance To provide a comprehensive review of inference methods in medical knowledge graphs that incorporate subgraph structures,and to offer valuable insights for future research in the field.Method/Process A thorough analysis of relevant literature is conducted,integrating knowledge of knowledge graphs and inference techniques.The study highlights the features and limitations of current leading models that use subgraph structures for knowledge inference,and compares them with various other inference methods.The advantages and disadvantages of these approaches are assessed in the context of specific domain tasks.Furthermore,the current state of applications and future prospects of subgraph-based inference in the medical domain are summarized.Result/Conclusion Future studies should aim to explore the interaction between diverse modalities of information in the medical field to enhance inference capabilities.This will contribute to the development of more comprehensive medical knowledge graphs,thereby providing effective solutions to clinical practical challenges.

关 键 词：知识推理 子图结构 图神经网络 知识图谱 

分 类 号：R-058[医药卫生]