基于电子病历的实体识别和知识图谱构建的研究-ChinaXiv.PDFVIP

/article/02-2019-12-046.html 计算机应用研究 基于电子病历的实体识别和知识图谱构建的研究* a, b a, b a, b 黄梦醒 ，李梦龙 ，韩惠蕊 (海南大学 a. 南海海洋资源利用国家重点实验室; b. 信息科学技术学院, 海口 570228) 摘 要：针对中文电子病历中命名实体识别和实体关系抽取研究方法中存在的问题，提出了一种基于双向长短时记忆 (bidirectional long short term memory) CRF(conditional random field) 网络 与 结合的实体识别和实体关系抽取方法。该方法 首先使用词嵌入技术将文本转换为数值向量，作为神经网络BiLSTM 的输入，再结合CRF 链式结构进行序列标注，输 出最大概率序列，并对识别结果知识图谱化。实验证明，该方法对中文电子病历进行实体识别和实体关系抽取时的准 F 确率、召回率、 值有明显的提升。实验结果满足临床中系统应用需求，对帮助研究构建临床决策支持系统、个性化 医疗推荐服务有引导作用。 关键词：实体识别；实体关系；长短时记忆网络；知识图谱 中图分类号：TP391.1 doi: 10.3969/j.issn.1001-3695.2018.07.0414 1 v Research on entity recognition and knowledge graph construction based on 6 electronic medical records 2 0 0 Huang Mengxinga, b, Li Menglonga, b, Han Huiruia, b 0 . (a. State Key Laboratory of Marine Resource Utilization in South China Sea, b. College of Information Science Technology 0 Hainan University, Haikou 570228, China) 1 8 1 Abstract: Aiming at the problems in the research methods of named entity recognition and entity relationship extraction in 0 Chinese electronic medical records) , this paper proposed an entity identification and entity relationship based on bidirectional 2 : long short term memory and conditional random field (CRF) . The method first used word embedding technology to convert v text into numerical vector, as the input of neural network BiLSTM, combined with CRF chain structure for sequence labeling, i X output the maximum probability sequence, and mapping th