基于Node2vec和知识注意力机制的诊断预测

doi:10.11896/jsjkx.210300070

摘要/Abstract

摘要： 诊断预测根据患者的历史健康状态预测未来的诊断信息,是个体化医疗决策的核心任务。电子健康记录是患者随时间推移的健康状况和临床护理的记录,它为诊断预测提供了丰富的纵向临床数据。然而,现有基于电子健康记录的诊断预测模型还不能完全了解隐藏的疾病进展模式;其次,细粒度诊断预测的性能很大程度上依赖于富含信息的特征。为了增强表达并改进学习,设计一种基于Node2vec和知识注意力的诊断预测模型。该模型基于Node2vec从医学本体的全局结构中捕捉潜在的医学知识并将诊断代码和分类代码映射为低维向量;利用分类代码嵌入向量对患者诊断的临床知识进行编码,进一步丰富患者细粒度健康状态的特征表示;设计一种知识注意力机制并与门控循环单元结合,将领域知识和电子健康记录进行融合,从患者历史健康状态中捕捉长期关联和疾病进展模式。在现实数据集上的实验结果表明,与最新方法相比,该模型显著地提高了预测性能。此外,结果表明Node2vec可以从医学本体捕捉到蕴含更多信息的医疗概念嵌入,知识注意力机制有助于促进外部知识和电子健康记录的有效融合。

关键词: Node2vec, 电子健康记录, 门控循环单元, 诊断预测, 知识注意力

Abstract: Diagnostic prediction predicts the future diagnosis of patients from their historical health states,and it is the core task of personalized medical decisions.Electronic health record(EHR) documents patients' time-varying health conditions and clinical care,and also provides a wealth of longitudinal clinical data for diagnostic prediction.However,the existing diagnostic prediction models based on EHR can not completely learn the hidden disease progression patterns.Moreover,the performance of fine-grained diagnostic prediction greatly depends on more informative sequence features.In order to improve the performance,we propose adiagnostic prediction model,called Node2vec and knowledge attention model (NKAM).Specifically,based on Node2vec,the model captures the potential medical knowledge from the global structure of medical ontology.It also maps categories into low-dimensional vectors and encodes the medical knowledge of patients' health state into category embedding vectors.The diagnosis code embedding vectorsare used to enrich the patients' fine-grained health state representation.Then,the long-term dependencies and disease progression patterns can be extracted from the patient's historical health states using a knowledge attention mechanism combined with the Gated Recurrent Unit(GRU).Experimental results on real-world dataset show that NKAM significantly improves the prediction performance compared with state-of-the-art methods.Furthermore,the experiments reveal that Node2vec can capture more informative medical concept embedding from medical ontology,and the knowledge-based attention mechanism helps to the effective integration of external knowledge and electronic health records.

Key words: Diagnostic prediction, Electronic health record(EHR), Gated recurrent unit(GRU), Knowledge attention, Node2vec

中图分类号:

TP391

李杭, 李维华, 陈伟, 杨仙明, 曾程. 基于Node2vec和知识注意力机制的诊断预测[J]. 计算机科学, 2021, 48(11A): 630-637. https://doi.org/10.11896/jsjkx.210300070

LI Hang, LI Wei-hua, CHEN Wei, YANG Xian-ming, ZENG Cheng. Diagnostic Prediction Based on Node2vec and Knowledge Attention Mechanisms[J]. Computer Science, 2021, 48(11A): 630-637. https://doi.org/10.11896/jsjkx.210300070

参考文献

[1]HÄYRINEN K,SARANTO K,NYKÄNEN P.Definition,structure,content,use and impacts of electronic health records:a review of the research literature[J].International Journal of Medical Informatics,2008,77(5):291-304.
[2]JENSEN P B,JENSEN L J,BRUNAK S.Mining electronichealth records:towards better research applications and clinical care[J].Nature Reviews Genetics,2012,13(6):395-405.
[3]YADAV P,STEINBACH M,KUMAR V,et al.Mining Elec-tronic Health Records (EHRs) A Survey[J].ACM Computing Surveys (CSUR),2018,50(6):1-40.
[4]SOLARES J R A,RAIMONDI F E D,ZHU Y,et al.Deep learning for electronic health records:A comparative review of multiple deep neural architectures[J].Journal of Biomedical Informatics,2020,101:103337.
[5]ABUL-HUSN N S,KENNY E E.Personalized medicine and the power of electronic health records[J].Cell,2019,177 (1):58-69.
[6]MURPHY K P.Book Machine learning:a probabilistic perspective [M] .Cambridge:MIT press,2012.
[7]SUN J,MCNAUGHTON C D,ZHANG P,et al.Predictingchanges in hypertension control using electronic health records from a chronic disease management program[J].Journal of the American Medical Informatics Association,2014,21(2):337-344.
[8]CHEN R,SU H,KHALILIA M,et al.Cloud-based predictive modeling system and its application to asthma readmission prediction[C]//AMIA Annual Symposium Proceedings.American Medical Informatics Association.2015:406.
[9]CHOI E,BAHADORI M T,SCHUETZ A,et al.Doctor ai:Predicting clinical events via recurrent neural networks[C]//Machine Learning for Healthcare Conference.2016:301-318.
[10]LEE W,PARK S,JOO W,et al.Diagnosis prediction via medical context attention networks using deep generative modeling[C]//2018 IEEE International Conference on Data Mining (ICDM).IEEE,2018:1104-1109.
[11]CHE Z,KALE D,LI W,et al.Deep computational phenotyping[C]//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2015:507-516.
[12]BAYTAS I M,XIAO C,ZHANG X,et al.Patient subtyping via time-aware lstm networks[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2017:65-74.
[13]NING S M,TENG F,LI T R.Entity relationship extraction of electronic medical records based on multi-channel self-attention mechanism[J].Chinese Journal of Computers,2020,43(5):916-929.
[14]CHOI E,BAHADORI M T,SONG L,et al.GRAM:graph-based attention model for healthcare representation learning[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2017:787-795.
[15]CHE Z,CHENG Y,ZHAI S,et al.Boosting deep learning risk prediction with generative adversarial networks for electronic health records[C]//2017 IEEE International Conference on Data Mining (ICDM).IEEE,2017:787-792.
[16]MA F,GAO J,SUO Q,et al.Risk prediction on electronic health records with prior medical knowledge[C]//Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining.2018:1910-1919.
[17]CHOI E,BAHADORI M T,SUN J,et al.Retain:An interpretable predictive model for healthcare using reverse time attention mechanism[C]//Advances in Neural Information Processing Systems.2016:3504-3512.
[18]MA F,CHITTA R,ZHOU J,et al.Dipole:Diagnosis prediction in healthcare via attention-based bidirectional recurrent neural networks[C]//Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2017:1903-1911.
[19]TROTT P.International classification of diseases for oncology[J].Journal of Clinical Pathology,1977,30 (8):782.
[20]COST H,PROJECT U.Clinical classifications software (CCS) for ICD-9-CM[OL].http://www hcup-us ahrq gov/toolssoftware/ccs/ccs jsp Accessed February,2018,27.
[21]STEARNS M Q,PRICE C,SPACKMAN K A,et al.SNOMED clinical terms:overview of the development process and project status[C]//Proceedings of the AMIA Symposium.American Medical Informatics Association.2001:662.
[22]BODENREIDER O.The unified medical language system(UMLS):integrating biomedical terminology[J].Nucleic acids research,2004,32 (suppl_1):D267-D270.
[23]LIPSCOMB C E.Medical subject headings (MeSH)[J].Bulletin of the Medical Library Association,2000,88(3):265.
[24]MA F,YOU Q,XIAO H,et al.Kame:Knowledge-based attention model for diagnosis prediction in healthcare[C]//Proceedings of the 27th ACM International Conference on Information and Knowledge Management.2018:743-752.
[25]TU C C,YANG C,LIU Z Y,et al.Overview of network representation learning[J].Scientia Sinica Informationis,2017,47(8):980-996.
[26]QI Z W,WANG J H,YUE K,et al.Graph embedding methods and applications:research review[J].Acta Electronica Sinica,2020,446(4):186-196.
[27]PEROZZI B,AL-RFOU R,SKIENA S.Deepwalk:Online learning of social representations[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2014:701-710.
[28]GROVER A,LESKOVEC J.node2vec:Scalable feature learning for networks[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2016:855-864.
[29]ZENG M,LI M,WU F X,et al.DeepEP:a deep learning framework for identifying essential proteins[J].BMC Bioinformatics,2019,20(16):506.
[30]ZENG M,LI M,FEI Z,et al.A deep learning framework for identifying essential proteins by integrating multiple types of biological information[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics,2019,18(1):296-305.
[31]BARABASI A L,OLTVAI Z N.Network biology:understanding the cell's functional organization[J].Nature Reviews Gene-tics,2004,5(2):101-113.
[32]BARABÁSI A L,GULBAHCE N,LOSCALZO J.Networkmedicine:a network-based approach to human disease[J].Nature Reviews Genetics,2011,12(1):56-68.
[33]MIKOLOV T,CHEN K,CORRADO G,et al.Efficient estimation of word representations in vector space[J].arXiv:13013781,2013.
[34]HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[35]CHO K,VAN MERRIËNBOER B,BAHDANAU D,et al.On the properties of neural machine translation:Encoder-decoder approaches[J].arXiv:14091259,2014.
[36]VASWANI A,SHAZEER N,PARMAR N,et al.Attention isall you need[C]//Advances in Neural Information Processing Systems.2017:5998-6008.
[37]JOHNSON A E,POLLARD T J,SHEN L,et al.MIMIC-III,a freely accessible critical care database[J].Scientific Data,2016,3(1):1-9.
[38]KINGMA D P,BA J.Adam:A method for stochastic optimization[J].arXiv:14126980,2014.
[39]SRIVASTAVA N,HINTON G,KRIZHEVSKY A,et al.Dropout:a simple way to prevent neural networks from overfitting[J].The Journal of Machine Learning Research,2014,15(1):1929-1958.
[40]LI M,ZHANG T,CHEN Y,et al.Efficient mini-batch training for stochastic optimization[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2014:661-670.

相关文章 14

[1]	赵丹丹, 黄德根, 孟佳娜, 董宇, 张攀. 基于BERT-GRU-ATT模型的中文实体关系分类 Chinese Entity Relations Classification Based on BERT-GRU-ATT 计算机科学, 2022, 49(6): 319-325. https://doi.org/10.11896/jsjkx.210600123
[2]	余皑欣, 冯秀芳, 孙静宇. 结合物品相似性的社交信任推荐算法 Social Trust Recommendation Algorithm Combining Item Similarity 计算机科学, 2022, 49(5): 144-151. https://doi.org/10.11896/jsjkx.210300217
[3]	刘文洋, 郭延哺, 李维华. 识别关键蛋白质的混合深度学习模型 Identifying Essential Proteins by Hybrid Deep Learning Model 计算机科学, 2021, 48(8): 240-245. https://doi.org/10.11896/jsjkx.200700076
[4]	后同佳, 周良. 基于双向GRU神经网络和注意力机制的中文船舶故障关系抽取方法 Chinese Ship Fault Relation Extraction Method Based on Bidirectional GRU Neural Network and Attention Mechanism 计算机科学, 2021, 48(11A): 154-158. https://doi.org/10.11896/jsjkx.210100215
[5]	郭鑫, 张庚, 陈千, 王素格. 面向机器阅读理解的候选句抽取算法 Candidate Sentences Extraction for Machine Reading Comprehension 计算机科学, 2020, 47(5): 198-203. https://doi.org/10.11896/jsjkx.190300154
[6]	屠袁飞,张成真. 面向云端的安全高效的电子健康记录 Secure and Efficient Electronic Health Records for Cloud 计算机科学, 2020, 47(2): 294-299. https://doi.org/10.11896/jsjkx.181202256
[7]	朱培培, 王中卿, 李寿山, 王红玲. 基于篇章信息和Bi-GRU的中文事件检测 Chinese Event Detection Based on Document Information and Bi-GRU 计算机科学, 2020, 47(12): 233-238. https://doi.org/10.11896/jsjkx.191100031
[8]	孙中锋, 王静. 用于基于方面情感分析的RCNN-BGRU-HN网络模型 RCNN-BGRU-HN Network Model for Aspect-based Sentiment Analysis 计算机科学, 2019, 46(9): 223-228. https://doi.org/10.11896/j.issn.1002-137X.2019.09.033
[9]	石春丹, 秦岭. 基于BGRU-CRF的中文命名实体识别方法 Chinese Named Entity Recognition Method Based on BGRU-CRF 计算机科学, 2019, 46(9): 237-242. https://doi.org/10.11896/j.issn.1002-137X.2019.09.035
[10]	何瑾琳, 刘学军, 徐新艳, 毛宇佳. 融合node2vec和深度神经网络的隐式反馈推荐模型 Implicit Feedback Recommendation Model Combining Node2vec and Deep Neural Networks 计算机科学, 2019, 46(6): 41-48. https://doi.org/10.11896/j.issn.1002-137X.2019.06.005
[11]	裴兰珍, 赵英俊, 王哲, 罗赟骞. 采用深度学习的DGA域名检测模型比较 Comparison of DGA Domain Detection Models Using Deep Learning 计算机科学, 2019, 46(5): 111-115. https://doi.org/10.11896/j.issn.1002-137X.2019.05.017
[12]	郑诚, 薛满意, 洪彤彤, 宋飞豹. 用于短文本分类的DC-BiGRU_CNN模型 DC-BiGRU_CNN Model for Short-text Classification 计算机科学, 2019, 46(11): 186-192. https://doi.org/10.11896/jsjkx.180901702
[13]	周枫, 李荣雨. 基于BGRU池的卷积神经网络文本分类模型 Convolutional Neural Network Model for Text Classification Based on BGRU Pooling 计算机科学, 2018, 45(6): 235-240. https://doi.org/10.11896/j.issn.1002-137X.2018.06.042
[14]	闫铭, 张应辉, 郑东, 吕柳迪, 苏昊楠. 灵活访问且模糊可搜索的EHR云服务系统 Flexibly Accessed and Vaguely Searchable EHR Cloud Service System 计算机科学, 2018, 45(10): 172-177. https://doi.org/10.11896／j.issn.1002-137X.2018.10.032

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed