摘要
现有大量慢阻肺患者跟电子病历,但偏远地区缺乏相关有经验的医生且患者缺乏对疾病的认识.知识图谱利于知识的展示,利于医生学习新的医学知识,也能普及患者对疾病的认识,因此本文提出一种构建慢阻肺知识图谱的方法及其中涉及到医学实体的命名实体识别问题的解决方法.首先对慢性阻塞性肺疾病诊治指南用自顶向下的方式设计Schema(概念)层,对中日友好医院的电子病历中的数据进行知识抽取,其中非结构化数据的知识抽取采用双向长短期记忆网络与条件随机场相结合的方法,通过设计实验,验证了该设计方法的准确性和有效性.
There are a large number of patients with chronic obstructive pulmonary disease(COPD)and electronic medical records,but there are no experienced doctors in remote areas and patients lack awareness of the disease.The knowledge map is conducive to the display of knowledge,and it is also beneficial for doctors to learn new medical knowledge,and also to popularize patients’understanding of diseases.Therefore,this paper proposes a method for constructing a knowledge map of chronic obstructive pulmonary disease and its name entity recognition problem involving medical entities.Solution.Firstly,the Schema layer is designed in a top-down manner for the diagnosis and treatment of chronic obstructive pulmonary disease,and the data in the electronic medical records of the China-Japan Friendship Hospital is extracted.The knowledge extraction of unstructured data adopts Bi-directional Long Short-Term Memory.The method of combining the memory network with the conditional random field has verified the accuracy and effectiveness of the design method through design experiments.
作者
贾辛洪
宋文爱
李伟岩
王青
雷毅
陈志华
常宗平
JIA Xin-hong;SONG Wen-ai;LI Wei-yan;WANG Qing;LEI Yi;CHEN Zhi-hua;CHANG Zong-ping(Software School,North University of China,Shanxi 030051,China;Research Institute of Information Technology,Tsinghua University,Beijing 100084,China;China-Japan Friendship Hospital,Institute of Clinical Medicine,Beijing 100084,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2020年第7期1371-1374,共4页
Journal of Chinese Computer Systems
基金
国家重点研发计划项目(2017YFC0910001)资助。
关键词
慢阻肺知识图谱
条件随机场
图数据库
命名实体识别
长短期记忆网络
chronic obstructive pulmonary disease knowledge graph
conditional random field(CRF)
named entity recognition
graph database
long short-term memory
