基于深度学习的电子病历命名实体识别的研究与实现

电子病历是医疗单位对门诊部、住院患者临床诊疗与指导干预的、数字化的医疗服务工作的相关记录1]。为了完成电子病历的高效的信息提取工作,本文使用深度学习的相关算法对电子病历中的文本进行命名实体的识别工作。其算法选择LSTM(Long-Short Term Memory,长短期记忆人工神经网络)和MLP(Multi-Layer Perception,多层神经网络),其用于构建算法模型。该本使用BP网络(Back-PropagationNetwork,后向传播)训练数据模型,应用已经标注的病历数据进行相应的训练与测试。该本通过实验证明,深度学习的算法在电子病历命名实体识别中是高效的2]。

Research and Implementation of Named Entity Recognition of Electronic Medical Records Based on Deep Learning

Electronic medical record is a digital medical service record of outpatient department and inpatient clinic diagnosis and intervention in medical units1]. To complete efficient information extraction of electronic medical records, the paper carries on text named entities recognition in electronic medical records with deep learning algorithm, including LSTM(Long-Short Term Memory) and MLP(Multi-Layer Perception) to construct algorithm model. Train data model with Back-Propagation Network(BP Network), and carry on corresponding train and test with labeled medical records. Experiment proves deep learning algorithm is efficient in electronic medical record named entity recognition2].