• 中国中文核心期刊
  • 中国科学引文数据库核心期刊
  • 中国科技核心期刊
  • 中国高校百佳科技期刊
高级检索

基于深度学习模型的我国药品不良反应报告实体关系抽取研究

陈瑶, 吴红, 葛卫红, 张海霞, 廖俊

陈瑶, 吴红, 葛卫红, 张海霞, 廖俊. 基于深度学习模型的我国药品不良反应报告实体关系抽取研究[J]. 中国药科大学学报, 2019, 50(6): 753-759. DOI: 10.11665/j.issn.1000-5048.20190617
引用本文: 陈瑶, 吴红, 葛卫红, 张海霞, 廖俊. 基于深度学习模型的我国药品不良反应报告实体关系抽取研究[J]. 中国药科大学学报, 2019, 50(6): 753-759. DOI: 10.11665/j.issn.1000-5048.20190617
CHEN Yao, WU Hong, GE Weihong, ZHANG Haixia, LIAO Jun. Research on entity relation extraction of Chinese adverse drug reaction reports based on deep learning method[J]. Journal of China Pharmaceutical University, 2019, 50(6): 753-759. DOI: 10.11665/j.issn.1000-5048.20190617
Citation: CHEN Yao, WU Hong, GE Weihong, ZHANG Haixia, LIAO Jun. Research on entity relation extraction of Chinese adverse drug reaction reports based on deep learning method[J]. Journal of China Pharmaceutical University, 2019, 50(6): 753-759. DOI: 10.11665/j.issn.1000-5048.20190617

基于深度学习模型的我国药品不良反应报告实体关系抽取研究

基金项目: 国家自然科学基金资助项目(No.81773806);双一流创新团队资助项目(No.CPU2018GY19);江苏省食品药品监督管理局2017—2018年度科研项目资助项目(No.20170308)

Research on entity relation extraction of Chinese adverse drug reaction reports based on deep learning method

  • 摘要: 药品不良反应(adverse drug reaction,ADR)报告作为药品上市后安全评价的主要载体,对药物安全评价研究具有重要的参考价值。本文以深度学习模型中的双向门控循环单元(bidirectional gated recurrent unit,Bi-GRU)结构为基础,引入注意力机制以及字向量与分词向量优化模型,对我国ADR报告中的ADR过程描述部分进行“药品-不良反应”的关系抽取研究。实验结果表明,基于深度学习的实体关系抽取模型在确认不良反应描述中“药品-不良反应”对之间的关系(否认、可能、直接和后处理)的分类任务中达到了很好的性能,最终模型取得87.52%的F值。所提取的信息在辅助ADR报告评价的同时,可进一步运用于特定药物的不良反应统计学研究以及知识库构建等任务中,从而为药物安全性评价研究提供更多的研究手段。
    Abstract: Adverse drug reaction(ADR)reports are acting as primary sources for post-marketing drug safety evaluation, which have important reference value for drug safety evaluation. In this article, bidirectional gated recurrent unit, a kind of deep learning method, was applied as the model for relation extraction of drugs and adverse reactions in free-text section of ADR descriptions in Chinese ADR reports, with attention as well as character embedding and word segmentation embedding added into the network. The experimental results showed that our model achieved good performance in the classification task of confirming the relationship of “Drug-ADR” pair(denial, likely, direct and post-therapy)in the ADR description, and the final model achieved an F-value of 87. 52%. The extracted information can assist in evaluating ADR reports and at the same time be utilized in tasks like statistical analysis of certain drugs and adverse events and ADR knowledge base construction to provide more research techniques for drug safety researches.
  • [1] Edwards IR, Aronson JK. Adverse drug reactions: definitions,diagnosis,and management[J].Lancet,2000,356(9237):1255-1259.
    [2] Chen Y, Zhou CJ, Li TX, et al. Named entity recognition from Chinese adverse drug event reports with lexical feature based BiLSTM-CRF and tri-training[J].J Biomed Inform,2019,96:103252.
    [3] Chen X,Liu X,Gales MJF,et al.Recurrent neural network language model training with noise contrastive estimation for speech recognition[C]//2015 IEEE International Conference on Acoustics,Speech and Signal Processing(ICASSP),2015.South Brisbane,Queensl and,Australia.New York,USA:IEEE,2015.
    [4] Wu WY,Chen YF,Xu JA,et al.Review of Chinese entity relation extraction[J],Comput mod,2018,(08):27-27,34.
    [5] Li MY,Yang J.Open chinese entity relation extraction method based on dependency parsing[J].Comput Eng,2016,42(06):201-207.
    [6] Luo X, Zhou WW, Wang WP, et al. Attention-based relation extraction with bidirectional gated recurrent unit and highway network in the analysis of geological data[J].IEEE Access,2018,6:5705-5715.
    [7] Ma JD,Liang LF,Xia CX.Entity relation extraction for free text in electronic medical records[J].J Med Intel,2016,37(12):1-9.
    [8] Liu K,Fu HD,Zou YW,et al.Chinese medical weak supervised relation extraction based on convolution neural network[J].Comput Sci,2017,44(10):249-253.
    [9] Kang N, Singh B, Bui C, et al. Knowledge-based extraction of adverse drug events from biomedical text[J].BMC Bioinform,2014,15:64.
    [10] Henriksson A,Kvist M,Dalianis H,et al.Identifying adverse drug event information in clinical notes with distributional semantic representations of context[J].J Biomed Inform,2015,57:333-349.
    [11] Liu J,Zhao SZ,Zhang XD.An ensemble method for extracting adverse drug events from social media[J].Artif Intell Med,2016,70:62-76.
    [12] Graves A.Supervised Sequence Labelling with Recurrent Neural Networks[J].Studies in Computational Intelligence,2008:385.
    [13] Cho K,van Merrienboer B,Gulcehre C,et al.Learning phrase representations using RNN Encoder-Decoder for statistical machine translation[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing(EMNLP),Doha,Qatar.Stroudsburg,PA,USA:Association for Computational Linguistics,2014.
    [14] Zhou P,Shi W,Tian J,et al.Attention-based bidirectional long short-term memory networks for relation classification[C]//Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics(Volume 2:Short Papers),Berlin,Germany.Stroudsburg,PA,USA:Association for Computational Linguistics,2016.
    [15] Huang DG,Jiang ZC,Zou L,et al.Drug-drug interaction extraction from biomedical literature using support vector machine and long short term memory networks[J].Inf Sci,2017,415/416:100-109.
    [16] Luo Y.Recurrent neural networks for classifying relations in clinical notes[J].J Biomed Inform,2017,72:85-95.
    [17] Kim J,Lee J.Multiple range-restricted bidirectional gated recurrent units with attention for relation classification[J].arXiv:Computation and Language,2017.
    [18] Xiao MG,Liu C.Semantic relation classification via hierarchical recurrent neural network with attention[C].//International Conference on Computational Linguistics,2016:1254-1263.
    [19] Chen X,Xu L,Liu Z,et al.Joint learning of character and word embeddings[C].//International Conference on Artificial Intelligence,2015:1236-1242.
    [20] Mikolov T,Sutskever I,Kai C,et al.Distributed representations of words and phrases and their compositionality[J].Advances in Neural Information Processing Systems,2013,26:3111-3119.
    [21] Lin Y, Shen S, Liu Z, et al. Neural relation extraction with selective attention over instances[C]//Meeting of the association for computational linguistics,2016:2124-2133.
  • 期刊类型引用(8)

    1. 吴菊华,李俊锋,陶雷. 基于知识图谱嵌入与深度学习的药物不良反应预测. 广东工业大学学报. 2024(01): 19-26+40 . 百度学术
    2. 任韡,黄彦,朱枫,喻锦扬,王青. 智能化药品不良反应报告辅助评价工具研究. 中国食品药品监管. 2024(09): 122-129 . 百度学术
    3. 荣丹琪,王倩,唐丽,司婉雨,赵鸿萍. 基于特征的深度学习预测化合物-蛋白质相互作用的研究进展. 中国药科大学学报. 2023(03): 305-313 . 本站查看
    4. 由丽萍,王世钰,李朝翻. 基于框架语义分析的社交网络药品不良事件抽取. 医学信息学杂志. 2023(07): 57-62 . 百度学术
    5. 卢恒,陈章杰,周知. 基于知识图谱的虚拟学术社区用户生成内容知识共聚框架研究. 情报理论与实践. 2023(12): 157-166+192 . 百度学术
    6. 仲雨乐,马诗雯,陆豪杰,韩普. 基于机器学习的药品不良反应实体识别研究综述. 软件工程. 2022(08): 1-6 . 百度学术
    7. 陈伟,吴云志,涂凌,刘航,余克健,乐毅. 基于多头自注意力机制的实体识别研究. 蚌埠学院学报. 2022(05): 54-60 . 百度学术
    8. 潘文洁,尹泽宇,侯婉馨,Jawad Hussain,张远鹏,姚敏,王理. 基于医学科研与临床应用的药物知识库研究进展. 中华临床医师杂志(电子版). 2021(01): 72-78 . 百度学术

    其他类型引用(3)

计量
  • 文章访问数:  697
  • HTML全文浏览量:  15
  • PDF下载量:  829
  • 被引次数: 11
出版历程
  • 刊出日期:  2019-12-24

目录

    /

    返回文章
    返回
    x 关闭 永久关闭