基于联合神经网络模型的中文医疗实体分类与关系抽取
|
摘要
近年来,医疗健康领域的实体分类与关系抽取引起了广泛关注。以往工作大多采用流水线模型,此类模型容易忽略任务间联系并造成错误传播,而联合学习则能够很好地避免这2个问题。为此,把卷积神经网络与支持向量机、条件随机场相结合,构建了联合神经网络模型。在此模型基础上,以参数共享的方式,分别通过任务联合、模型联合以及特征联合对实体分类与关系抽取2个任务进行联合学习,在药品说明书语料库中取得了非常不错的效果,实体分类和关系抽取的F值分别达到了98.0%和98.3%。实验表明,联合神经网络模型对于实体分类和关系抽取是非常有效的。
Entity classification and relationship extraction in healthcare and medical field have attracted wide attention in recent years, and most of the work used the pipeline model in the past, which easily ignored the link between tasks and caused error propagation. Since joint learning can well avoid the two problems, a joint neural network model is established by combining the convolution neural network with support vector machine and conditional random field. On the basis of this model, entity classification and relationship extraction are studied jointly by task combination, model combination and feature combination in the way of parameter sharing. The joint neural network model achieves very good performance in the medicine instructions corpus, and the F-scores of entity classification and relationship extraction reach 98.0% and 98.3%, respectively. Experiments show that the joint neural network model is very effective for entity classification and relationship extraction in the medicine instructions corpus.
Entity classification and relationship extraction in healthcare and medical field have attracted wide attention in recent years, and most of the work used the pipeline model in the past, which easily ignored the link between tasks and caused error propagation. Since joint learning can well avoid the two problems, a joint neural network model is established by combining the convolution neural network with support vector machine and conditional random field. On the basis of this model, entity classification and relationship extraction are studied jointly by task combination, model combination and feature combination in the way of parameter sharing. The joint neural network model achieves very good performance in the medicine instructions corpus, and the F-scores of entity classification and relationship extraction reach 98.0% and 98.3%, respectively. Experiments show that the joint neural network model is very effective for entity classification and relationship extraction in the medicine instructions corpus.
引文
[1] Hanisch D,Fundel K,Mevissen H T,et al.ProMiner:Rule-based protein and gene entity recognition[J].BMC bioinformatics,2005,6(1):S14-S14.
[2] Khoo C S G,Chan S,Niu Y.Extracting causal knowledge from a medical database using graphical patterns[C]//Proc of the 38th Annual Meeting on Association for Computational Linguistics,2000:336-343.
[3] Dingare S,Nissim M,Finkel J,et al.A system for identifying named entities in biomedical text:How results from two evaluations reflect on both the system and the evaluations[J].International Journal of Genomics,2005,6(1-2):77-85.
[4] Frunza O,Inkpen D.Extraction of disease-treatment semantic relations from biomedical sentences[C]//Proc of the 2010 Workshop on Biomedical Natural Language Processing,2010:91-98.
[5] Chen E S,Hripcsak G,Xu H,et al.Automated acquisition of disease drug knowledge from biomedical and clinical documents:An initial study[J].American Medical Informatics Association,2008,15(1):87-98.
[6] Tang Nan,Yang Zhi-hao,Lin Hong-fei,et al.Protein-protein interaction extraction from medical literature based on multiple kernels learning[J].Computer Engineering,2011,37(10):184-186.(in Chinese)
[7] Yang Jin-feng,Yu Qiu-bin,Guan Yi,et al.An overview of research on electronic medical record oriented named entity recognition and entity relation extraction[J].Acta Automatica Sinica,2014,40(8):1537-1562.(in Chinese)
[8] Yang Jin-feng,Guan Yi,He Bin,et al.Corpus construction for named entities and entity relations on Chinese electronic medical records[J].Journal of Software,2016,27(11):2725-2746.(in Chinese)
[9] Liu M,Jiang L,Hu H.Automatic extraction and visualization of semantic relations between medical entities from medicine instructions[J].Multimedia Tools and Applications,2017,76(8):10555-10573.
[10] Wu Y,Zhao J,Xu B,et al.Chinese named entity recognition based on multiple features[C]//Proc of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing,2005:427-434.
[11] Jagannatha A N,Yu H.Bidirectional RNN for medical event detection in electronic health records[C]//Proc of NAACL-HLT 2016,2016:473-482.
[12] Zheng S,Hao Y,Lu D,et al.Joint entity and relation extraction based on a hybrid neural network[J].Neurocomputing,2017,257:59-66.
[13] Adel H,Schütze H.Global normalization of convolutional neural networks for joint entity and relation classification[J].arXiv preprint arXiv:1707.07719,2017.
[14] Roth D,Yih W.A linear programming formulation for global inference in natural language tasks[C]//Proc of the 8th Conference on Computational Natural Language Learning(CoNLL-2004) at HLT-NAACL 2004,2004:1-8.
[15] Zheng S,Wang F,Bao H,et al.Joint extraction of entities and relations based on a novel tagging scheme[J].arXiv preprint arXiv:1706.05075,2017.
[16] Li F,Zhang M,Fu G,et al.A neural joint model for entity and relation extraction from biomedical text[J].BMC Bioinformatics,2017,18(1):198-198.
[17] Gurulingappa H,Rajput A M,Roberts A,et al.Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports[J].Journal of Biomedical Informatics,2012,45(5):885-892.
[18] Lample G,Ballesteros M,Subramanian S,et al.Neural architectures for named entity recognition[J].arXiv preprint arXiv:1603.01360,2016.
[19] Nitsuwat S,Paoin W.Development of ICD-10-TM ontology for a semi-automated morbidity coding system in Thailand[J].Methods of Information in Medicine,2012,51(6):519-528.
[6] 唐楠,杨志豪,林鸿飞,等.基于多核学习的医学文献蛋白质与科学关系抽取[J].计算机工程,2011,37(10):184-186.
[7] 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.
[8] 杨锦锋,关毅,何彬,等.中文电子病历命名实体和实体关系语料库构建[J].软件学报,2016,27(11):2725-2746.
[2] Khoo C S G,Chan S,Niu Y.Extracting causal knowledge from a medical database using graphical patterns[C]//Proc of the 38th Annual Meeting on Association for Computational Linguistics,2000:336-343.
[3] Dingare S,Nissim M,Finkel J,et al.A system for identifying named entities in biomedical text:How results from two evaluations reflect on both the system and the evaluations[J].International Journal of Genomics,2005,6(1-2):77-85.
[4] Frunza O,Inkpen D.Extraction of disease-treatment semantic relations from biomedical sentences[C]//Proc of the 2010 Workshop on Biomedical Natural Language Processing,2010:91-98.
[5] Chen E S,Hripcsak G,Xu H,et al.Automated acquisition of disease drug knowledge from biomedical and clinical documents:An initial study[J].American Medical Informatics Association,2008,15(1):87-98.
[6] Tang Nan,Yang Zhi-hao,Lin Hong-fei,et al.Protein-protein interaction extraction from medical literature based on multiple kernels learning[J].Computer Engineering,2011,37(10):184-186.(in Chinese)
[7] Yang Jin-feng,Yu Qiu-bin,Guan Yi,et al.An overview of research on electronic medical record oriented named entity recognition and entity relation extraction[J].Acta Automatica Sinica,2014,40(8):1537-1562.(in Chinese)
[8] Yang Jin-feng,Guan Yi,He Bin,et al.Corpus construction for named entities and entity relations on Chinese electronic medical records[J].Journal of Software,2016,27(11):2725-2746.(in Chinese)
[9] Liu M,Jiang L,Hu H.Automatic extraction and visualization of semantic relations between medical entities from medicine instructions[J].Multimedia Tools and Applications,2017,76(8):10555-10573.
[10] Wu Y,Zhao J,Xu B,et al.Chinese named entity recognition based on multiple features[C]//Proc of the Conference on Human Language Technology and Empirical Methods in Natural Language Processing,2005:427-434.
[11] Jagannatha A N,Yu H.Bidirectional RNN for medical event detection in electronic health records[C]//Proc of NAACL-HLT 2016,2016:473-482.
[12] Zheng S,Hao Y,Lu D,et al.Joint entity and relation extraction based on a hybrid neural network[J].Neurocomputing,2017,257:59-66.
[13] Adel H,Schütze H.Global normalization of convolutional neural networks for joint entity and relation classification[J].arXiv preprint arXiv:1707.07719,2017.
[14] Roth D,Yih W.A linear programming formulation for global inference in natural language tasks[C]//Proc of the 8th Conference on Computational Natural Language Learning(CoNLL-2004) at HLT-NAACL 2004,2004:1-8.
[15] Zheng S,Wang F,Bao H,et al.Joint extraction of entities and relations based on a novel tagging scheme[J].arXiv preprint arXiv:1706.05075,2017.
[16] Li F,Zhang M,Fu G,et al.A neural joint model for entity and relation extraction from biomedical text[J].BMC Bioinformatics,2017,18(1):198-198.
[17] Gurulingappa H,Rajput A M,Roberts A,et al.Development of a benchmark corpus to support the automatic extraction of drug-related adverse effects from medical case reports[J].Journal of Biomedical Informatics,2012,45(5):885-892.
[18] Lample G,Ballesteros M,Subramanian S,et al.Neural architectures for named entity recognition[J].arXiv preprint arXiv:1603.01360,2016.
[19] Nitsuwat S,Paoin W.Development of ICD-10-TM ontology for a semi-automated morbidity coding system in Thailand[J].Methods of Information in Medicine,2012,51(6):519-528.
[6] 唐楠,杨志豪,林鸿飞,等.基于多核学习的医学文献蛋白质与科学关系抽取[J].计算机工程,2011,37(10):184-186.
[7] 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.
[8] 杨锦锋,关毅,何彬,等.中文电子病历命名实体和实体关系语料库构建[J].软件学报,2016,27(11):2725-2746.