Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations

Tamás Dózsa; Gergö Bognar; Peter Kovacs

doi:10.1007/978-3-030-45096-0_44

Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations

Tamás Dózsa, Gergö Bognar, Peter Kovacs

Institute of Signal Processing

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

Abstract

In this work, we are focusing on the problem of heartbeat classification in electrocardiogram (ECG) signals. First we develop a patient-specific feature extraction scheme by using adaptive orthogonal transformations based on wavelets, B-splines, Hermite and rational functions. The so-called variable projection provides the general framework to find the optimal nonlinear parameters of these transformations. After extracting the features, we train a support vector machine (SVM) for each model whose outputs are combined via ensemble learning techniques. In the experiments, we achieved an accuracy of 94.2% on the PhysioNet MIT-BIH Arrhythmia Database that shows the potential of the proposed signal models in arrhythmia detection.

Original language	English
Title of host publication	Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS)
Editors	Roberto Moreno-Díaz, Alexis Quesada-Arencibia, Franz Pichler
Publisher	Springer
Pages	355-363
Number of pages	9
Volume	12014
ISBN (Print)	978-3-030-45096-0
DOIs	https://doi.org/10.1007/978-3-030-45096-0_44
Publication status	Published - Apr 2020

Publication series

Name	Lecture Notes in Computer Science (LNCS)

Fields of science

102019 Machine learning
202 Electrical Engineering, Electronics, Information Engineering
202022 Information technology
202037 Signal processing

JKU Focus areas

Digital Transformation

Access to Document

10.1007/978-3-030-45096-0_44

Cite this

Dózsa, T., Bognar, G., & Kovacs, P. (2020). Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations. In R. Moreno-Díaz, A. Quesada-Arencibia, & F. Pichler (Eds.), Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS) (Vol. 12014, pp. 355-363). (Lecture Notes in Computer Science (LNCS)). Springer. https://doi.org/10.1007/978-3-030-45096-0_44

Dózsa, Tamás ; Bognar, Gergö ; Kovacs, Peter. / Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations. Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS). editor / Roberto Moreno-Díaz ; Alexis Quesada-Arencibia ; Franz Pichler. Vol. 12014 Springer, 2020. pp. 355-363 (Lecture Notes in Computer Science (LNCS)).

@inproceedings{359c89ee1ea241fcb3003e36ef786a3a,

title = "Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations",

abstract = "In this work, we are focusing on the problem of heartbeat classification in electrocardiogram (ECG) signals. First we develop a patient-specific feature extraction scheme by using adaptive orthogonal transformations based on wavelets, B-splines, Hermite and rational functions. The so-called variable projection provides the general framework to find the optimal nonlinear parameters of these transformations. After extracting the features, we train a support vector machine (SVM) for each model whose outputs are combined via ensemble learning techniques. In the experiments, we achieved an accuracy of 94.2\% on the PhysioNet MIT-BIH Arrhythmia Database that shows the potential of the proposed signal models in arrhythmia detection.",

author = "Tam{\'a}s D{\'o}zsa and Gerg{\"o} Bognar and Peter Kovacs",

year = "2020",

month = apr,

doi = "10.1007/978-3-030-45096-0\_44",

language = "English",

isbn = "978-3-030-45096-0",

volume = "12014",

series = "Lecture Notes in Computer Science (LNCS)",

publisher = "Springer",

pages = "355--363",

editor = "Roberto Moreno-D{\'i}az and Alexis Quesada-Arencibia and Franz Pichler",

booktitle = "Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS)",

}

Dózsa, T, Bognar, G & Kovacs, P 2020, Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations. in R Moreno-Díaz, A Quesada-Arencibia & F Pichler (eds), Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS). vol. 12014, Lecture Notes in Computer Science (LNCS), Springer, pp. 355-363. https://doi.org/10.1007/978-3-030-45096-0_44

Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations. / Dózsa, Tamás; Bognar, Gergö; Kovacs, Peter.
Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS). ed. / Roberto Moreno-Díaz; Alexis Quesada-Arencibia; Franz Pichler. Vol. 12014 Springer, 2020. p. 355-363 (Lecture Notes in Computer Science (LNCS)).

Research output: Chapter in Book/Report/Conference proceeding › Conference proceedings › peer-review

TY - GEN

T1 - Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations

AU - Dózsa, Tamás

AU - Bognar, Gergö

AU - Kovacs, Peter

PY - 2020/4

Y1 - 2020/4

N2 - In this work, we are focusing on the problem of heartbeat classification in electrocardiogram (ECG) signals. First we develop a patient-specific feature extraction scheme by using adaptive orthogonal transformations based on wavelets, B-splines, Hermite and rational functions. The so-called variable projection provides the general framework to find the optimal nonlinear parameters of these transformations. After extracting the features, we train a support vector machine (SVM) for each model whose outputs are combined via ensemble learning techniques. In the experiments, we achieved an accuracy of 94.2% on the PhysioNet MIT-BIH Arrhythmia Database that shows the potential of the proposed signal models in arrhythmia detection.

AB - In this work, we are focusing on the problem of heartbeat classification in electrocardiogram (ECG) signals. First we develop a patient-specific feature extraction scheme by using adaptive orthogonal transformations based on wavelets, B-splines, Hermite and rational functions. The so-called variable projection provides the general framework to find the optimal nonlinear parameters of these transformations. After extracting the features, we train a support vector machine (SVM) for each model whose outputs are combined via ensemble learning techniques. In the experiments, we achieved an accuracy of 94.2% on the PhysioNet MIT-BIH Arrhythmia Database that shows the potential of the proposed signal models in arrhythmia detection.

UR - https://www.springer.com/gp/book/9783030450953

U2 - 10.1007/978-3-030-45096-0_44

DO - 10.1007/978-3-030-45096-0_44

M3 - Conference proceedings

SN - 978-3-030-45096-0

VL - 12014

T3 - Lecture Notes in Computer Science (LNCS)

SP - 355

EP - 363

BT - Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS)

A2 - Moreno-Díaz, Roberto

A2 - Quesada-Arencibia, Alexis

A2 - Pichler, Franz

PB - Springer

ER -

Dózsa T, Bognar G, Kovacs P. Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations. In Moreno-Díaz R, Quesada-Arencibia A, Pichler F, editors, Computer Aided Systems Theory - EUROCAST 2019, Part II, Lecture Notes in Computer Science (LNCS). Vol. 12014. Springer. 2020. p. 355-363. (Lecture Notes in Computer Science (LNCS)). doi: 10.1007/978-3-030-45096-0_44

Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations

Abstract

Publication series

Fields of science

JKU Focus areas

Access to Document

Other files and links

Morphology based Cardiac Signal Processing

Cite this

Ensemble Learning for Heartbeat Classification Using Adaptive Orthogonal Transformations

Abstract

Publication series

Fields of science

JKU Focus areas

Access to Document

Other files and links

Projects

Morphology based Cardiac Signal Processing

Cite this