摘要
阐述基于BiLSTM-CRF基准模型,融合词向量及词属性推理机制实现中文电子病历命名实体识别的方法,包括医学语料库构建与词向量训练、融合词属性推理机制等,分析实验结果,指出医学领域词向量及词属性推理机制的引入有助于提升中文电子病历命名实体识别效果。
The paper introduces the method of combining word vector and word attribute reasoning to realize Named Entity Recognition(NER) of Chinese Electronic Medical Record(EMR) based on BiLSTM-CRF benchmark model,including medical corpus construction,word vector training,integration of word attribute reasoning,etc.,analyzes the experimental results,and points out that the introduction of word vector and word attribute reasoning in medical field can improve NER effect of Chinese EMR.
作者
武学鸿
杨峰
李建华
徐倩
WU Xuehong;YANG Feng;LI Jianhua;XU Qian(School of Computer Science and Engineering,Central South University,Changsha 410083;Research Institute,Hunan Creator Information Technology Co.Ltd.,Changsha 410205;Xiangya School of Medicine,Central South University,Changsha 410003)
出处
《医学信息学杂志》
CAS
2022年第7期39-42,共4页
Journal of Medical Informatics
基金
国际科技创新合作基地项目“湖南省人工智能与医学大数据国际科技创新合作基地”(项目编号:2019CB1007)。
关键词
领域词向量
词属性推理机制
中文电子病历
命名实体识别
自然语言处理
domain word vector
word attribute reasoning
Chinese Electronic Medical Record(EMR)
Named Entity Recognition(NER)
Natural Language Processing(NLP)
