基于标签共现和特征局部相关的心电异常检测方法

doi:10.11896/jsjkx.220200004

计算机科学 ›› 2023, Vol. 50 ›› Issue (3): 139-146.doi: 10.11896/jsjkx.220200004

• 数据库&大数据&数据科学 • 上一篇下一篇

基于标签共现和特征局部相关的心电异常检测方法

韩京宇, 钱龙, 葛康, 毛毅

南京邮电大学计算机学院南京 210023
江苏省大数据安全与智能处理重点实验室南京 210023

收稿日期:2022-01-29 修回日期:2022-05-12 出版日期:2023-03-15 发布日期:2023-03-15
通讯作者: 韩京宇(jyhan@njupt.edu.cn)
基金资助:
国家自然科学基金(62002174)

ECG Abnormality Detection Based on Label Co-occurrence and Feature Local Pertinence

HAN Jingyu, QIAN Long, GE Kang, MAO Yi

School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China
Jiangsu Key Laboratory of Big Data Security & Intelligent Processing,Nanjing 210023,China

Received:2022-01-29 Revised:2022-05-12 Online:2023-03-15 Published:2023-03-15
About author:HAN Jingyu,born in 1976,Ph.D,professor,is a member of China Computer Federation.His main research interests include biomedical data processing,machine learning and spatio-temporal database.
Supported by:
National Natural Science Foundation of China(62002174).

摘要/Abstract

摘要： 自动的心电异常识别是一个多标签分类问题,多通过对每个标签训练一个二分类器来实现异常识别。由于异常数目多,特征和异常间以及不同异常间的相关性复杂,自动检测的效果并不理想。为了充分利用异常和特征间的依存关系,提出了一种基于异常标签共现和特征局部相关(Label Co-occurrence and Feature's local Pertinence,LCFP)的心电异常识别方法。首先,根据标签共现性和特征局部相关性,为标签构建包含宏特征和微特征的联合特征空间。宏特征采用狄利克雷过程混合模型聚类构建,以区分不同的共现标签集;微特征是原始特征空间的一个子集,用于区分共现标签集中的各个标签。进而,在联合特征空间为每个异常训练一个一对多(One-Versus-All)的概率分类器。其次,为充分利用异常的关联,提出在概率分类器排序基础上区分相关和非相关标签,采用Beta分布自适应地学习锚阈值和相关度阈值,以确定实例的相关标签集。LCFP是一种检测多种心电异常的通用方法,提高了心电异常识别的精度。在两个真实数据集上,F1指标分别提高了4%和22.4%,验证了所提方法的有效性。

关键词: 心电异常, 多标签分类, 标签共现, 狄利克雷过程混合模型, Beta分布, 锚阈值

Abstract: Automatic electrocardiogram(ECG) abnormality detection is a multi-label classification problem,which is commonly solved by training a binary-relevance classifier for each abnormality.Due to the large number of abnormalities,the complex correlations between features and abnormalities,and those among different abnormalities,existing methods’ performance is not satis-fying.To make full use of the dependencies between features and abnormalities,this paper proposes a novel abnormality detection method based on label co-occurrence and feature local pertinence(LCFP).Firstly,we set up a consolidated feature space consisting of both the macro-features and micro-features based on the label co-occurrence and features’ pertineance.The macro-features are constructed with a clustering approach based on Dirichlet process mixture model(DPMM),thus distinguishing differentco-occurrence label sets.The micro-features are a subset of primitive features,which serves to distinguish between the labels in the same labelset.Next,we train a one-versus-all classifier which returns a relevance probability.Secondly,to make use of the diffe-rent correlation degrees among different abnormalities,we propose to differ the relevant labels from the irrelevant ones based on the sorting according to the probabilities given by the classifiers.In particular,we propose to exploit the Beta distribution to adaptively learn the anchor thresholds and correlation thresholds,thus determining the relevant labels of an instance.Our LCFP me-thod is a universal way to detect every possible ECG abnormalities,which effectively improves the detection accuracy.The results on two real datasets show that our method can achieves an improvement of 4% and 22.4%,respectively,in terms of F1,which proves the effectiveness of our method.

Key words: Electrocardiogram abnormality, Multi-label classification, Label co-occurrence, Dirichlet process mixture model, Beta distribution, Anchor thresholds

中图分类号:

TP311.132

韩京宇, 钱龙, 葛康, 毛毅. 基于标签共现和特征局部相关的心电异常检测方法[J]. 计算机科学, 2023, 50(3): 139-146. https://doi.org/10.11896/jsjkx.220200004

HAN Jingyu, QIAN Long, GE Kang, MAO Yi. ECG Abnormality Detection Based on Label Co-occurrence and Feature Local Pertinence[J]. Computer Science, 2023, 50(3): 139-146. https://doi.org/10.11896/jsjkx.220200004

参考文献

[1]WORLD H O.Cardio-vascular diseases(CVDs) [EB/OL].(2021-06-11) [2021-06-11].https://www.who.int/en/news-room/ fact-sheets/detail/cardiovascular-diseases-(cvds).
[2]YANG H.Course book of Electrocardiogram Specialty[M].Beijing:Beijing University Medical Press,2005:18-34.
[3]INAN G,ELIF D U.ECG beat classifier designed by combined neural network model[J].Pattern Recognition,2005,38(2):199-2008.
[4]ZHU W L,CHEN X H,WANG Y,et al.Arrhythmia Recognition and Classification Using ECG Morphology and Segment Feature Analysis[J].IEEE/ACM Transactions on Computational Biology and Bioinformatics,2019,16(1):131-138.
[5]YANG C D,JIA Z,LI X W.Study on ECG Signal Recognition and Classification Based on U-Net++[J].Computer Science,2021,48(10):121-126.
[6]JIN L P,DONG J.Deep learning research on clinical electrocardiogram analysis [J].SCIENTIA SINICA Informationis,2015,45(3):398-416.
[7]ZHANG M L,ZHOU Z H.A Review on Multi-Label Learning Algorithms[J].IEEE Transactions on Knowledge and Data Engineering,2014,26(8):1819-1837.
[8]EVA G,SEBASTIAN V.A Tutorial on Multilabel Learning[J].ACM Computing Surveys,2015,47(3):1-38.
[9]TAE J J,HYUN J P,YOUNG-HAK K.Premature ventricular contraction beat detection with deep neuralnetworks[C]//Proceedings of 15th IEEE International Conference on Machine Learning and Applications.New York:IEEE Press,2016:49-56.
[10]TAE J J,HOANG M N,DAEYOUN K,et al.ECG arrhythmia classification using a 2-D convolutional neural network[EB/OL].(2018-03-24) [2018-04-25].https://dblp.org/rec/bib/journals/corr/abs-1804-06812.
[11]TOMÁS T,PAULO F,JESÚS P.Heartbeat ClassificationUsing Abstract Features From the Abductive interpretation of the ECG[J].IEEE Journal of Biomedical and Health Informa-tics,2018,22(2):409-420.
[12]HARI M R,ANURAG T,SHAILJA S.ECG signal processing for abnormalities detection using multi-resolution wavelet transform and Artificial Neural Network classifier [J].Measurement,2013,46(9):3238-3246.
[13]WANG Z,LI X F,ZHU H J,et al.Impact of Left Bundle Branch Area Pacing on Clinical and Electrocardiogram Characteristics Among Bradycardia Patients With Right Bundle Branch Block [J].Chinese Circulation Journal,2021,36(1):22-27.
[14]CELIN S,VASANTH K.ECG Signal Classific-ation Using Various Machine Learning Techniques[J].Journal of Medical Systems,2018,42(2018):240-250.
[15]LI K,DU N,ZHANG A D.Detecting ECG Abnormalities via Transductive Transfer learning[C]//Proceedings of ACM-BCB.New York:IEEE Press,2012:35-74.
[16]KEIJI G,SHOTA H,HIDEH-IRO O,et al.Building NormalECG Models to Detect Any Arrhythmias Using Deep Learning[C]//Proceedings of the 47th International Conference on Computing in Cardiology.New York:IEEE Press,2020:45-51.
[17]LI F,MIAO D Q,ZHANG Z F,et al.Mutual information-based granular feature-weighted multi-label learning k-nearest neighbor algorithm[J].Computer Research and Development,2017,54(5):1024-1035.
[18]MATTHEW R,BOUTELL,LUO J B,et al.Learning multi-label scene classification[J].Pattern Recognition,2004,37(9):1757-1771.
[19]JESSE R,BERNHARD P,GEOFF H,et al.Classifier chains for multi-label classification[C]//ECML PKDD 2009.Berlin:Springer,2009:254-269.
[20]JOHANNES F,EYKE H,ENELDO L M,et al.Multi-label classification via calibrated label ranking[J].Machine Learning,2008,73:133-153.
[21]GRIGORIOS T,IOANNIS K,IOANNIS V.Random k-labelsets for multi-label classification [J].IEEE Transactions on Know-ledge and Data Engineering,2011,23(7):1079-1089.
[22]ZHANG M L,ZHOU Z H.ML-KNN:A Lazy Learning Ap-proach to Multi-Label Learning[J].Pattern Recognition,2007,40(7):2038-2048.
[23]AMANDA C,ROSS D.KING.Knowledge discovery in multi-label phenotype data[C]//Proceedings of International Confe-rence on Principles of Data Mining and Knowledge Discovery.2001,42-53.
[24]ELISSEEFF A,WESTON J.A kernel method for multi-labelled classifica-tion[C]//Proceedings of the 14th International Conference on Neural Information Processing Systems:Natural and Synthetic.New York:IEEE Press,2001:681-687.
[25]NADIA G,ANDREW M C.Collective Multi-Label Classification[C]//Proceedings of the 14th ACM International Conference on Information and Knowledge Managemen.Berlin:Springer Press,2005:93-99.
[26]HALL M.Correlation-Based feature selection for discrete andnumeric class machine learning[C]//Proceedings of the 17th International Conference on Machine Learning.Berlin:Springer Press,2000:359-366.
[27]RON K,GEORGE H.John.Wrappers for Feature Subset Selection [J].Artifical Intelligence,1997,97(1):273-324.
[28]ZHANG M L,WU L.LIFT:Multi-Label Learning with Label-Specific Features [J].IEEE Transactions on Pattern Analysis and Machine Learning,2015,37(1):107-120.
[29]HAMED D H,MARYAM M.A Multi Rate Marginalized Parti-cle Extended Kalman Filter for P and T Wave Segmentation in ECG Signals[J].IEEE Journal of Biomedical and Health Informatics,2019,23(1):112-122.
[30]LI C W,ZHENG C X,TAI C F.Detection of ECG Characteristic Points Using Wavelet Transforms[J].IEEE Transactions on Biomedical Engineering,1995,42(1):21-28.
[31]SMIT H W,VERTON K,GRIMBERGEN C A.A Low-CostMultichannel Preamplifier for Physiological Signals[J].IEEE Transactions on Biomedical Engineering,1987,34(4):307-310.
[32]LAGUNA P,THAKOR N V.New algorithm for QT intervalanalysis in 24 hour Hotler ECG:Performance and applications [J].Medical Biological Engineering and Computing,1990,28:67-73.
[33]SALAH H,ASMA B A,MOHAMED H B.A robust QRS complex detection using regular grammar and deterministic automata [J].Biomedical Signal Processing and Control,2018,40:263-274.
[34]HAN J W,MICHELINE K,PEI J.Data Mining:Concepts and Techniques [M].AMSTERDAM:Morgan Kaufmann,2012:443-474.
[35]FERGUSON T S.A Bayesian Analysis of Some Non-Parametric Problems [J].The Annals of Statistics,1973,1(2):209-230.
[36]Masked ECGDATASET [EB/OL].(2020-03-01)[2021-10-27].https://github.com/hjyresearch228/ PAC.
[37]MIT-BIHARRHYTHMIAS [EB/OL].(2005-03-01)[2005-03-27].https://physionet.org/content/m-itdb/ 1.0.0/.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于标签共现和特征局部相关的心电异常检测方法

ECG Abnormality Detection Based on Label Co-occurrence and Feature Local Pertinence

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 6

Metrics

本文评价

推荐阅读 0

[1]	武红鑫, 韩萌, 陈志强, 张喜龙, 李慕航. 监督和半监督学习下的多标签分类综述 Survey of Multi-label Classification Based on Supervised and Semi-supervised Learning 计算机科学, 2022, 49(8): 12-25. https://doi.org/10.11896/jsjkx.210700111
[2]	孙开伟, 郭豪, 曾雅苑, 方阳, 刘期烈. 一种基于超网络的多目标回归方法 Multi-target Regression Method Based on Hypernetwork 计算机科学, 2022, 49(11A): 211000205-9. https://doi.org/10.11896/jsjkx.211000205
[3]	李玲,刘华文,徐晓丹,赵建民. 基于信息增益的多标签特征选择算法 Multi-label Feature Selection Algorithm Based on Information Gain 计算机科学, 2015, 42(7): 52-56. https://doi.org/10.11896/j.issn.1002-137X.2015.07.012
[4]	王磊,黄河笑,吴兵,郑任儿. 基于主题与三支决策的文本情感分析 Emotion Analysis of Text Based on Topics and Three-way Decisions 计算机科学, 2015, 42(6): 93-96. https://doi.org/10.11896/j.issn.1002-137X.2015.06.021
[5]	王磊,苗夺谦,张志飞,余鹰. 基于主题的文本句情感分析 Emotion Analysis on Text Sentences Based on Topics 计算机科学, 2014, 41(3): 32-35.
[6]	刘涛,熊焰,黄文超,陆琦玮,关亚文. 一种基于Bayes估计的WSN节点信任度计算模型 Trust Computation Model of Nodes Based on Bayes Estimation in Wireless Sensor Networks 计算机科学, 2013, 40(10): 61-64.