A novel multi-class classification method for arrhythmias using Hankel dynamic mode decomposition and long short-term memory networks

Chuchu Liang; Majid Khan Majahar Ali; Lili Wu

doi:10.46481/jnsps.2025.2411

Authors

Chuchu Liang School of Mathematical Sciences, Universiti Sains Malaysia, Pulau Pinang, 11800, Malaysia https://orcid.org/0009-0002-3668-0602
Majid Khan Majahar Ali
[email protected]
School of Mathematical Sciences, Universiti Sains Malaysia, Pulau Pinang, 11800, Malaysia https://orcid.org/0000-0002-5558-5929
Lili Wu School of Mathematical Sciences, Universiti Sains Malaysia, Pulau Pinang, 11800, Malaysia

Keywords:

Dynamic Mode Decomposition (DMD), Long Short-Term Memory (LSTM), Arrhythmia, Multi-Class

Abstract

The complex dynamic properties of ECG signals and the challenge of multi-category classification make automated diagnosis of arrhythmias difficult. In this paper, we propose a new model for the multiclassification task of arrhythmia, which combines Hankel Dynamic Modal Decomposition (HDMD) and Long Short-Term Memory Network (LSTM).HDMD is used to construct the Hankel matrix, the optimal delay parameter is selected based on 90% of the energy of the singular values, and dynamic modal features extracted from it are used as the input sequences of LSTM. The LSTM model is optimisation is performed by minimising the cross-entropy loss function, setting the maximum number of iterations to 60 and using an early stopping strategy to avoid overfitting. The model was validated on the MIT-BIH arrhythmia database, which contains 109,402 beats and is classified into five categories according to the AAMI criteria: normal beats (N), ventricular premature beats (V), fusion beats (F), atrial premature beats (S) and unclassifiable (Q). By comparing with the direct use of LSTM, the experimental results showed that the HDMD[1]LSTM model showed different degrees of improvement in all five classifications, especially in the classification of atrial premature beats (S) and ventricular premature beats (V), and the overall classification accuracy improved to 0.85. Future work can focus on two aspects: first, to solve the category imbalance problem, by over sampling or undersampling techniques to improve the classification ability of a few categories; second, exploring the embedding of DMD into the network structure of LSTM to further optimise the feature extraction and classification performance.

Dimensions

REFERENCES

[1] C. Antzelevitch & A. Burashnikov, “Overview of basic mechanisms of cardiac arrhythmia”, Cardiac Electrophysiology Clinics 3 (2011) 23. https://doi.org/10.1016/j.ccep.2010.10.012.

[2] R. Gopinathannair, S. P. Etheridge, F. E. Marchlinski, F. G. Spinale, D. Lakkireddy & B. Olshansky, “Arrhythmia-induced cardiomyopathies: mechanisms, recognition, and management”, Journal of the American College of Cardiology 66 (2015) 1714. https://doi.org/10.1016/j.jacc.2015.08.038.

[3] M. Saber & M. Abotaleb, “Arrhythmia modern classification techniques: A review”, Journal of Artificial Intelligence and Metaheuristics 1 (2022) 42. https://doi.org/10.54216/JAIM.010205.

[4] P. J. Schmid, “Dynamic mode decomposition and its variants”, Annual Review of Fluid Mechanics 54 (2022) 225. https://doi.org/10.1146/annurev-fluid-030121-015835.

[5] H. N. Ingabire, K. Wu, J. T. Amos, S. He, X. Peng, W. Wang & P. Ren, “Analysis of ECG signals by dynamic mode decomposition”, IEEE Journal of Biomedical and Health Informatics 26 (2021) 2124. https://doi.org/10.1109/JBHI.2021.3130275.

[6] Y. Yu, X. Si, C. Hu & J. Zhang, “A review of recurrent neural networks: LSTM cells and network architectures”, Neural Computation 31 (2019) 1235. https://doi.org/10.1162/neco_a_01199.

[7] M. G. Tsipouras, D. I. Fotiadis & D. Sideris, “An arrhythmia classification system based on the RR-interval signal”, Artificial Intelligence in Medicine 33 (2005) 237. https://doi.org/10.1016/j.artmed.2004.03.007.

[8] Y. Kutlu, G. Altan & N. Allahverdi, “Arrhythmia classification using waveform ECG signals”, presented at International Conference on Advanced Technology & Sciences, Konya, Turkey, 2016. https://www.researchgate.net/publication/308919881_Arrhythmia_Classification_Using_Waveform_ECG_Signals.

[9] J. A. Nasiri, M. Naghibzadeh, H. S. Yazdi & B. Naghibzadeh, “ECG arrhythmia classification with support vector machines and genetic algorithm”, Proceedings of the 2009 Third UKSim European Symposium on Computer Modeling and Simulation, 2009, pp. 187--192. https://doi.org/10.1109/EMS.2009.39.

[10] B. M. Asl, S. K. Setarehdan & M. Mohebbi, “Support vector machine based arrhythmia classification using reduced features of heart rate variability signal”, Artificial Intelligence in Medicine 44 (2008) 51. https://doi.org/10.1016/j.artmed.2008.04.007.

[11] W. H. Jung and S. G. Lee, “An arrhythmia classification method utilizing the weighted KNN and the fitness rule”, IRBM 38 (2017) 138. https://doi.org/10.1016/j.irbm.2017.04.002.

[12] W. M. Zuo, W. G. Lu, K. Q. Wang & H. Zhang, “Diagnosis of cardiac arrhythmia using kernel difference weighted KNN classifier”, 2008 Computers in Cardiology (2008) 253. https://doi.org/10.1109/CIC.2008.4749025.

[13] R. Kumari Mohebbanaaz, L. V. Rajani & Y. Padma Sai, “Classification of arrhythmia beats using optimized K-nearest neighbor classifier”, Intelligent Systems: Proceedings of ICMIB (2020) 349–359. https://doi.org/10.1007/978-981-33-6081-5_31.

[14] Z. Ebrahimi, M. Loni, M. Daneshtalab & A. Gharehbaghi, “A review on deep learning methods for ECG arrhythmia classification”, Expert Systems with Applications 7 (2020) 100033. https://doi.org/10.1016/j.eswax.2020.100033.

[15] C. Chen, Z. Hua, R. Zhang, G. Liu & W. Wen, “Automated arrhythmia classification based on a combination network of CNN and LSTM”, Biomedical Signal Processing and Control 57 (2020) 101819. https://doi.org/10.1016/j.bspc.2019.101819.

[16] T. J. Jun, H. M. Nguyen, D. Kang, D. Kim, D. Kim & Y. H. Kim, “ECG arrhythmia classification using a 2-D convolutional neural network”, arXiv preprint (2018). https://www.semanticscholar.org/paper/ECG-arrhythmia-classification-using-a-2-D-neural-Jun-Nguyen/622d18e3622f05bfe2054703318ea71747975dd0.

[17] M. Salem, S. Taheri & J. S. Yuan, “ECG arrhythmia classification using transfer learning from 2-dimensional deep CNN features”, presented at the 2018 IEEE Biomedical Circuits and Systems Conference, Cleveland, OH, USA, 2018, pp. 1--4. https://doi.org/10.1109/BIOCAS.2018.8584808.

[18] B. Hou, J. Yang, P. Wang & R. Yan, “LSTM-based auto-encoder model for ECG arrhythmias classification”, IEEE Transactions on Instrumentation and Measurement 69 (2019) 1232. https://doi.org/10.1109/TIM.2019.2910342.

[19] P. Liu, X. Sun, Y. Han, Z. He, W. Zhang & C. Wu, “Arrhythmia classification of LSTM autoencoder based on time series anomaly detection”, Biomedical Signal Processing and Control 71 (2022) 103228. https://doi.org/10.1016/j.bspc.2021.103228.

[20] C. Chen, Z. Hua, R. Zhang, G. Liu & W. Wen, “Automated arrhythmia classification based on a combination network of CNN and LSTM”, Biomedical Signal Processing and Control 57 (2020) 101819. https://doi.org/10.1016/j.bspc.2019.101819.

[21] R. He, Y. Liu, K. Wang, N. Zhao, Y. Yuan, Q. Li & H. Zhang, “Automatic cardiac arrhythmia classification using combination of deep residual network and bidirectional LSTM”, IEEE Access 7 (2019) 102119. https://doi.org/10.1109/ACCESS.2019.2931500.

[22] E. Vasconcelos Filho and P. L. dos Santos, “A dynamic mode decomposition approach with Hankel blocks to forecast multi-channel temporal series”, IEEE Control Systems Letters 3 (2019) 739. https://doi.org/10.1109/LCSYS.2019.2917811.

[23] J. H. Tu, Dynamic mode decomposition: Theory and applications, Ph.D. dissertation, Princeton University, 2013. https://robotics.caltech.edu/wiki/images/c/c7/DMDApplicationsTheory.pdf.

[24] J. Chung, C. Gulcehre, K. Cho & Y. Bengio, “Empirical evaluation of gated recurrent neural networks on sequence modeling”, arXiv preprint, arXiv:1412.3555 (2014). https://arxiv.org/abs/1412.3555.

[25] Y. LeCun, Y. Bengio & G. Hinton, “Deep learning”, Nature 521 (2015) 436. https://doi.org/10.1038/nature14539.

[26] A. Goldberger, L. Amaral, L. Glass, J. Hausdorff, P. C. Ivanov, R. Mark & H. E. Stanley, “PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals”, Circulation 101 (2000) e215. https://doi.org/10.1161/01.CIR.101.23.e215.

[27] B. Zhao, H. Liu, T. Bi & S. Xu, “Synchrophasor measurement method based on cascaded infinite impulse response and dual finite impulse response filters”, Journal of Modern Power Systems and Clean Energy (2024). https://doi.org/10.35833/MPCE.2023.000824.

[28] F. Mumtaz, T. S. Khan, M. Alqahtani, H. A. Sher, A. S. Aljumah & S. Z Almutairi, “Ultra high-speed fault diagnosis scheme for DC distribution systems based on discrete median filter and mathematical morphology”, IEEE Access (2024). https://doi.org/10.1109/ACCESS.2024.3381519.

[29] P. De Chazal, Detection of supraventricular and ventricular ectopic beats using a single lead ECG, Proceedings of the 2013 35th Annual Interna tional Conference of the IEEE Engineering in Medicine and Biology Society (EMBC), IEEE, 2013, pp. 45–48. https://doi.org/10.1109/EMBC.2013.6609433.

[30] T. Pokaprakarn, R. R. Kitzmiller, J. R. Moorman, D. E. Lake, A. K. Krishnamurthy & M. R. Kosorok, “Sequence to sequence ECG cardiac rhythm classification using convolutional recurrent neural networks”, IEEE Journal of Biomedical and Health Informatics 26 (2021) 572. https://doi.org/10.1109/JBHI.2021.3098662.

[31] X. Wang & L. Sun, “Extracting dynamic mobility patterns by Hankel dynamic modes decomposition”, presented at The 11th Triennial Symposium on Transportation Analysis, Mauritius Island, Apr. 2022.

[32] L. Fu, B. Lu, B. Nie, Z. Peng, H. Liu & X. Pi, “Hybrid network with attention mechanism for detection and location of myocardial infarction based on 12-lead electrocardiogram signals”, Sensors 20 (2020) 1020. https://doi.org/10.3390/s20041020.

[33] J. Hu, W. Zhao, D. Jia, C. Yan, H. Wang, Z. Li, J. Fang & M. Yang, “Deep multi-instance networks for bundle branch block detection from multi lead ECG”, in Proc. 2020 42nd Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. (EMBC), 2020, pp. 20–24. https://doi.org/10.1109/EMBC44109.2020.9175909.

[34] L. Guo, G. Sim & B. Matuszewski, “Inter-patient ECG classification with convolutional and recurrent neural networks”, Biocybernetics and Biomedical Engineering 39 (2019) 868. https://doi.org/10.1016/j.bbe.2019.06.001.

[35] Y. Gan, J. C. Shi, W. M. He & F. J. Sun, “Parallel classification model of arrhythmia based on DenseNet-BiLSTM”, Biocybernetics and Biomedical Engineering 41 (2021) 1548. https://doi.org/10.1016/j.bbe.2021.09.001.

[36] A. Chen, F. Wang, W. Liu, S. Chang, H. Wang, J. He & Q. Huang, “Multiinformation fusion neural networks for arrhythmia automatic detection”, Comput. Methods Programs Biomed 193 (2020) 105479. https://doi.org/10.1016/j.cmpb.2020.105479.

[37] E. Essa and X. Xie, “An ensemble of deep learning-based multi-model for ECG heartbeats arrhythmia classification”, IEEE Access 9 (2021) 103452. https://doi.org/10.1109/ACCESS.2021.3098986.

[38] Q. Xie, S. Tu, G. Wang, Y. Lian & L. Xu, “Feature enrichment based convolutional neural network for heartbeat classification from electrocardiogram”, IEEE Access 7 (2019) 153751. https://doi.org/10.1109/ACCESS.2019.2948857.

[39] J. Zhu, J. Lv & D. Kong, “CNN-FWS: A model for the diagnosis of normal and abnormal ECG with feature adaptive”, Entropy 24 (2022) 471. https://doi.org/10.3390/e24040471.

[40] S. Ran, X. Yang, M. Liu, Y. Zhang, C. Cheng, H. Zhu & Y. Yuan, “Homecare-oriented ECG diagnosis with large-scale deep neural network for continuous monitoring on embedded devices”, IEEE Transactions on Instrumentation and Measurement 71 (2022) 1. https://doi.org/10.1109/TIM.2022.3147328.

[41] J. He, J. Rong, L. Sun, H. Wang & Y. Zhang, “An advanced two step DNN-based framework for arrhythmia detection”, in Advances in Knowledge Discovery and Data Mining, Springer International Publishing, 2020, pp. 422–434. https://doi.org/10.1007/978-3-030-47436-2 32.