基于多源健康感知数据动静态关系融合的疾病诊断

doi:10.11896/jsjkx.211100241

摘要/Abstract

摘要： 疾病诊断是电子健康记录数据挖掘的热门研究领域,也是实现医疗诊断智能化的一个重要环节。但是,电子健康记录中健康感知数据的来源多样、数据结构复杂,且不同类型的数据之间有着潜在的相关性,在进行特征提取和挖掘分析过程中存在着异构数据应该如何融合的问题。只有对医学感测数据、个人体质记录数据、疾病间关系数据进行综合考虑,挖掘其中的相关隐藏特征,才能对多种类别疾病进行更准确的诊断。因此,基于多源健康感知数据动静态关系融合的疾病诊断模型(DSRF)首先通过动静态关系融合算法解决动态医学感测数据和静态体质记录数据的异构性问题并挖掘其相关关系,然后计算多类别疾病的关联矩阵来提取疾病间依赖关系,最后在门控循环单元网络架构的基础上将多种健康感知数据进行融合,完成了多源异构数据的综合分析。在美国MIMIC-III临床数据集上的实验结果证明,相比同类型主流模型,该模型可以更准确地对多种类别疾病进行联合诊断。

关键词: 多源数据融合, 动静态关系融合, 疾病诊断, 电子健康记录, 临床数据挖掘

Abstract: Disease diagnosis is a field of electronic health record data mining where lots of researchers are interested in,and it is also an important link to realize the intellectualization of medical diagnosis.However,due to the diversity of data sources,complex data structure and potential correlation among different types of health sensing data,there is a problem of how to fuse heterogeneous data in the process of feature extraction and data mining.Therefore,comprehensively considering clinical sensing data,personal physical record data and relationship data between diseases,and mining the latent relevant features can make the diagnosis of multi-category diseases more accurate.Dynamic and static relationship fusion of multi-source health perception data for disease diagnosis(DSRF) is proposed.Firstly,the dynamic and static relationship fusion algorithm is used to extract data correlation features and solve the heterogeneity of dynamic clinical sensing time series data and static personal physical condition data.Then the dependency matrix of multi-category diseases is calculated to extract the correlations among diseases.Finally,various health sen-sing data is fused based on the gated recurrent unit network.The comprehensive analysis of multi-source heterogeneous data is completed after the above three steps.Experimental results on the real-world American MIMIC-III clinical dataset show that the proposed model outperforms state-of-the-art models and is able to diagnose multiple categories of diseases accurately.

Key words: Multi-source data fusion, Dynamic and static relationship fusion, Disease diagnosis, Electronic health record, Clinical data mining

中图分类号:

TP399

霍甜媛, 顾晶晶. 基于多源健康感知数据动静态关系融合的疾病诊断[J]. 计算机科学, 2022, 49(11A): 211100241-9. https://doi.org/10.11896/jsjkx.211100241

HUO Tian-yuan, GU Jing-jing. Dynamic and Static Relationship Fusion of Multi-source Health Perception Data for Disease Diagnosis[J]. Computer Science, 2022, 49(11A): 211100241-9. https://doi.org/10.11896/jsjkx.211100241

参考文献

[1]KOREN A,PRASAD R.IoT Health Data in Electronic Health Records(EHR):Security and Privacy Issues in Era of 6G[J].Journal of ICT Standardization,2022,10(1):63-84.
[2]MAHAJAN P,RANA D.Investigating Clinical Named EntityRecognition Approaches for Information Extraction from EMR[M]//Tracking and Preventing Diseases with Artificial Intelligence.Springer,Cham,2022:153-175.
[3]MOHAMMADI R,JAIN S,AGBOOLA S,et al.Learning toidentify patients at risk of uncontrolled hypertension using electronic health records data[J].AMIA Summits on Translational Science Proceedings,2019,2019:533-542.
[4]MA L,GAO J,WANG Y,et al.AdaCare:Explainable Clinical Health Status Representation Learning via Scale-Adaptive Feature Extraction and Recalibration[C]//Proceedings of the AAAI Conference on Artificial Intelligence,2020:825-832.
[5]AYON S I,ISLAM M M.Diabetes Prediction:A Deep LearningApproach[J].International Journal of Information Engineering and Electronic Business,2019,11(2):21-27.
[6]JOHNSON A E W,POLLARD T J,SHEN L,et al.MIMIC-III,a freely accessible critical care database[J].Scientific Data,2016,3(1):1-9.
[7]WONG M S,WELLS M,PARRINELLA K,et al.EHR phenotyping by Natural Language Processing improves detection of patients at risk for preeclampsia[J].American Journal of Obstetrics & Gynecology,2022,226(1):S65-S66.
[8]LIAO B,JIA X,ZHANG T,et al.DHDIP:An InterpretableModel for Hypertension and Hyperlipidemia Prediction Based on EMR Data[J/OL].SSRN.http://dx.doi.org/10.2139/ssrn.4022954.
[9]GUTIERREZ G.Artificial intelligence in the intensive care unit[J].Critial Care,2020,24(1):1-9.
[10]PIRRACCHIO R.Mortality prediction in the ICU based onMIMIC-II results from the super ICU learner algorithm(SICULA) project[J/OL].Secondary Analysis of Electronic Health Records,2016,2016:295-313.https://doi.org/10.1007/978-3-3-19-43742-2_20.
[11]SANJAY P,CHUIZHENG M,ZHENGPING C,et al.Benchmarking deep learning models on large healthcare datasets[J].Journal of Biomedical Informatics,2018,83:112-134.
[12]HARUTYUNYAN H,KHACHATRIAN H,KALE D C,et al.Multitask learning and benchmarking with clinical time series data[J].Scientific Data,2019,6(1):1-18.
[13]LEE W,SHI Y,SUN H,et al.MSIPA:Multi-Scale Interval Pattern-Aware Network for ICU Transfer Prediction[J].ACM Transactions on Knowledge Discovery from Data(TKDD),2021,16(1):1-17.
[14]CHOI E,XU Z,LI Y,et al.Learning the Graphical Structure of Electronic Health Records with Graph Convolutional Transformer[J].Proceedings of the AAAI Conference on Artificial Intelligence,2020,34(1):606-613.
[15]XU Z,SO D R,DAI A M.MUFASA:Multimodal Fusion Architecture Search for Electronic Health Records[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2021:10532-10540.
[16]AGARWAL V,PODCHIYSKA T,BANDA J M,et al.Learning statistical models of phenotypes using noisy labeled training data[J].Journal of the American Medical Informatics Association,2016,23(6):1166-1173.
[17]HALPERN Y,HORNG S,CHOI Y,et al.Electronic medical record phenotyping using the anchor and learn framework[J].Journal of the American Medical Informatics Association,2016,23(4):731-740.
[18]MARLIN B M,KALE D C,KHEMANI R G,et al.Unsuper-vised pattern discovery in electronic health care data using proba-bilistic clustering models[C]//Proceedings of the 2nd ACM SIGHIT International Health Informatics Symposium.2012:389-398.
[19]HO J C,GHOSH J,SUN J.Marble:high-throughput phenotyping from electronic health records via sparse nonnegative tensor factorization[C]//Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2014:115-124.
[20]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.
[21]CHOI E,BAHADORI M T,SCHUETZ A,et al.Doctor ai:Predicting clinical events via recurrent neural networks[C]//Machine Learning for Healthcare Conference.PMLR,2016:301-318.
[22]RAZAVIAN N,MARCUS J,SONTAG D.Multi-task prediction of disease onsets from longitudinal laboratory tests[C]//Machine Learning for Healthcare Conference.PMLR,2016:73-100.
[23]MA L,ZHANG C,WANG Y,et al.Concare:Personalized clinical feature embedding via capturing the healthcare context[C]//Proceedings of the AAAI Conference on Artificial Intelligence.2020:833-840.
[24]CHUNG J,GULCEHRE C,CHO K H,et al.Empirical evaluation of gated recurrent neural networks on sequence modeling[J].arXiv:1412.3555,2014.
[25]LIPTON Z C,KALE D C,ELKAN C,et al.Learning to diag-nose with LSTM recurrent neural networks[J].arXiv:1511.03677,2015.
[26]HOCHREITER S,SCHMIDHUBER J.Long short-term memory[J].Neural Computation,1997,9(8):1735-1780.
[27]CHOI E,BAHADORI M T,SEARLES E,et al.Multi-layer representation learning for medical concepts[C]//proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.2016:1495-1504.
[28]ESTEBAN C,STAECK O,BAIER S,et al.Predicting clinicalevents by combining static and dynamic information using recurrent neural networks[C]//2016 IEEE International Conference on Healthcare Informatics(ICHI).IEEE,2016:93-101.
[29]SONG H,RAJAN D,THIAGARAJAN J J,et al.Attendand diagnose:Clinical time series analysis using attention models[C]//Thirty-second AAAI Conference on Artificial Intelligence.2018.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed