摘要
知识图谱提供的服务质量在很大程度上取决于知识图谱构建的质量.自动构建知识图谱的方法已被广泛应用于许多领域,但知识图谱在医学领域的应用却面临着很多困难,原因有:医学概念/关系/事件的复杂和模糊性;数据标准不一致,源数据质量差;医疗数据异构多元化严重,如电子医学病例(electronic medical record,简称EMR)等.在构建过程中,需要来自医学专家的大量先验知识和人工辅助.引入一个系统架构,该架构明确了在何时何处引入医学专家的相关工作,从而提高医疗健康知识图谱构建的质量和效率.
The quality of service (QoS) that a knowledge graph can provide largely depends on the quality of the knowledge. Automatic methods have been widely used in many domains to construct the knowledge graphs. However, it is more complex and difficult in the medical domain. There are three reasons: the complex and obscure nature of medical concepts and relations, inconsistent standards and heterogeneous multi-source medical data with low quality like EMRs. Therefore, the quality of knowledge graph requires a lot of manual efforts from experts in the process. In this paper, we introduced an overall framework that provided insights on where and when to import manual efforts in the process to construct a health knowledge graph. In this framework, four tools were provided to facilitate the doctors’ contribution, i.e. matching synonym, discovering and editing new concepts, annotating concepts and relations, together with establishing rule base. The application for cardiovascular diseases demonstrated that this framework could improve the accuracy and efficiency of medical knowledge graph construction.
作者
盛明
张勇
邢春晓
SHENG Ming;ZHANG Yong;XING Chunxiao(Web and Software R&D Center of Research Institute of Information Technology,Tsinghua University,Beijing 100084,China)
出处
《安徽大学学报(自然科学版)》
CAS
北大核心
2019年第6期48-54,共7页
Journal of Anhui University(Natural Science Edition)
基金
国家自然科学基金资助项目(91646202)
国家重点研发计划基金资助项目(2018YFB1404400,2018YFB1402700).
关键词
医疗知识图谱构建
医在回路
电子医学病例
medical knowledge graph construction
doctor-in-the-loop
EMR (electronic medical record)