医学知识图谱自动构建研究

Automatic Construction of Medical Knowledge Graph

随着我国医疗信息化水平的发展,大量医疗数据迅速产生,对如此丰富的医学知识进行有效利用显得尤为重要。医学知识图谱提供了对医学知识进行合理组织、管理和利用的手段。本文首先介绍了知识图谱的定义及架构,之后从实体抽取、关系抽取和实体对齐三个方面梳理了医学知识图谱自动构建过程的常用技术,分析了该领域的研究现状及最新技术进展,总结了医学知识图谱在医疗语义搜索引擎、医疗问答系统及医疗决策支持系统方面的应用现状,并分析了医学知识图谱自动构建及应用方面面临的挑战,旨在为医学知识图谱的构建及应用提供借鉴。

With the development of our country’s medical informatization level,a large amount of medical data has been rapidly generated,and it is particularly important to effectively use such a wealth of medical knowledge.The medical knowledge graph provides a means to organize,manage and utilize medical knowledge.This article first introduces the definition and architecture of the knowledge graph,and then introduces the common techniques in the medical knowledge graph automatic construction steps including entity extraction,relationship extraction and entity alignment.The research status of knowledge graph automatic construction techniques is summarized,and the latest research progress is analyzed.Finally,the application status of medical knowledge graph in medical semantic search engines,medical question answering systems and medical decision support systems are introduced,the challenges faced in the automatic construction techniques and applications of medical knowledge graph are analyzed,in order to provides a reference for the construction and application of medical knowledge graph.