[目的/意义]在线医疗信息抽取是实现医疗信息检索、医疗信息推荐、个人医疗健康提醒及警示、疾病诊断、公众健康监控、药物不良反应挖掘等服务的基础环节，而医疗实体抽取则是在线医疗信息抽取的首要工作。本文拟解决传统医疗实体抽取严重依赖于人工特征提取且效率低的问题。[方法/过程]以网络文本为研究对象，首先对医疗实体类型和医疗实体抽取的目标进行描述。将在线医疗文本中的医疗实体抽取任务看作序列标注问题来解决，通过对CNN模型和BiLSTM模型基础理论的探讨，构建基于混合深度学习模型CNN-BiLSTM的医疗实体抽取框架。[结果/结论]通过三组对比实验，验证了本文所使用的CNN-BiLSTM模型在医疗实体抽取任务中的有效性。
[Purpose/significance] Online medical information extraction is the basic link to achieve medical information retrieval, medical information recommendation, personal medical health reminder and warning, disease diagnosis, public health monitoring, drug adverse reaction mining and other services, while medical entity extraction is the primary work of online medical information extraction. This paper aims to solve the problem that traditional medical entity extraction relies heavily on artificial feature extraction and the problem of low efficiency.[Method/process] Taking network text as the research object, this paper firstly describes the type of medical entity and the target of extraction of medical entity. Online entity extraction task in medical text was considered a sequence labeling problem to solve, the paper discussed the basic theories of BiLSTM model and the CNN model, and built a model based on hybrid deep learning CNN-BiLSTM medical entity extraction framework.[Result/conclusion] At last, the effectiveness of the CNN-BiLSTM model in the medical entity extraction task was verified through three comparison experiments.
[1] 程健一. 中文电子病历实体关系抽取研究[D]. 哈尔滨:哈尔滨工业大学, 2016.
[2] WEI Q, CHEN T, XU R, et al. Disease named entity recognition by combining conditional random fields and bidirectional recurrent neural networks[J]. Database:the journal of biological databases & curation, 2016(1):1-8.
[3] 杨锦锋,于秋滨,关毅,等. 电子病历命名实体识别和实体关系抽取研究综述[J]. 自动化学报, 2014, 40(8):1537-1562.
[4] 龙光宇, 徐云. CRF与词典相结合的疾病命名实体识别[J]. 微型机与应用, 2017(21):51-53.
[5] LIU H, QIN X, FU B. The symptoms and pathogenesis entity recognition of TCM medical records based on CRF[C]//Proceedings of the 2017 international workshop on big data and information security. New York:IEEE, 2016:1479-1484.
[6] 苏娅, 刘杰, 黄亚楼. 在线医疗文本中的实体识别研究[J]. 北京大学学报(自然科学版), 2016, 52(1):1-9.
[7] CHIU J P C, NICHOLS E. Named entity recognition with bidirectional LSTM-CNNs[J]. Transactions of the Association for Computational Linguistics, 2016, 4(1):357-370.
[8] 杨红梅, 李琳, 杨日东,等. 基于双向LSTM神经网络电子病历命名实体的识别模型[J]. 中国组织工程研究, 2018(20):3237-3242.
[9] 吴嘉伟, 关毅, 吕新波. 基于深度学习的电子病历中实体关系抽取[J]. 智能计算机与应用, 2014, 4(3):35-38.
[10] XU K, ZHOU Z, HAO T, et al. A bidirectional LSTM and conditional random fields approach to medical named entity recognition[C]//Proceedings of the international conference on advanced intelligent systems and informatics 2017. Berlin:Springer, 2018:355-365.
[11] DONG X, CHOWDHURY S, QIAN L, et al. Transfer bi-directional LSTM RNN for named entity recognition in Chinese electronic medical records[C]//Proceedings of the international conference on E-Health networking, applications and services. New York:IEEE, 2017:1-4.
[12] MIOTTO R, WANG F, WANG S, et al. Deep learning for healthcare:review, opportunities and challenges[J]. Briefings in bioinformatics, 2017,19(6):1236-1246.
[13] i2b2. 2010i2b2/VA challenge evaluation concept annotation guidelines[EB/OL].[2018-02-18].https://www.i2b2.org/NLP/Relations/assets/Concept%20Annotation%20Guideline.pdf.
[14] KOUTNIK J, GREFF K, GOMEZ F, et al. A clockwork RNN[J]. Computer science, 2014:1863-1871.
[15] VU N T, ADEL H, GUPTA P, et al. Combining recurrent and convolutional neural networks for relation classification[C]//Proceedings of NAACL-HLT. Stroudsburg:ACL, 2016:534-539.
[16] WEN Y, ZHANG W, LUO R, et al. Learning text representation using recurrent convolutional neural network with highway layers[C]//Proceedings of Neu-IR'6 SIGIR workshop on neural information retrieval. New York:ACM, 2016:1-5.
[17] ADEL H, SCHUTZE H. Exploring different dimensions of attention for uncertainty detection[C]//Proceedings of the 15th conference of the European chapter of the Association for Computational Linguistics. Stroudsburg:ACL, 2017(1):22-34.
[18] YIN W, SCHUTZE H, XIANG B, et al. ABCNN:attention-based convolutional neural network for modeling sentence pairs[J]. Transactions of the Association for Computational Linguistics, 2016(4):259-272.
[19] DAUPHIN Y N, FAN A, AULI M, et al. Language modeling with gated convolutional networks[C]//Proceedings of the 34th international conference on machine learning. Sydney:The Journal of Machine Learning Research, 2017:1-9.
[20] ARKHIPENKO K, KOZLOV I, TROFIMOVICH J, et al. Comparison of neural network architectures for sentiment analysis of Russian Tweets[C]//Proceedings of the international conference "Dialogue 2016". Moscow:Computational Linguistics and Intellectual Technologies, 2016:1-10.
[21] HE K, ZHANG X, REN S, et al. Delving deep into rectifiers:Surpassing human-level performance on imagenet classification[C]//Proceedings of the IEEE international conference on computer vision. New York:IEEE, 2015:1026-1034.
[22] SANTOS C N D, ZADROZNY B. Learning character-level representations for part-of-speech tagging[C]//Proceedings of international conference on machine learning. Beijing:JMLR, 2014:1818-1826.
[23] GRAVES A, SCHMIDHUBER J. Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Netw, 2005, 18(5):602-610.
[24] RAMSHAW L A, MARCUS M P. Text Chunking using transformation-based learning[J]. Text speech & language technology, 1995, 11(5):82-94.
[25] GRAVES A, MOHAMED A R, HINTON G. Speech recognition with deep recurrent neural networks[C]//Proceedings of IEEE international conference on acoustics, speech and signal processing. New York:IEEE, 2013:6645-6649.
