A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Identifying patients with Atrial Fibrillation (AFib) is one of the most challenging and prevailing problems in cardiology. In this study, we propose a novel feature extraction method hybridizing probabilistic symbolic pattern recognition (PSPR) and Sample Entropy (SampEn) to represent morphological changes in electrocardiogram (ECG) recordings. We implement a PSPR framework on continuous SampEn and RR interval series obtained from 4,630 ECG recordings in the training dataset. In our hybrid feature extraction method, PSPR symbolically discretizes SampEn and RR interval series with seven and nine unique symbols, respectively and then models the pattern transition behavior of these series using probability theory. We extract 28 features including PSPR-based metrics and descriptive metrics from SampEn, RR intervals, and processed ECG recordings. A random-forest classifier was trained on 13 features derived using a Genetic Algorithm based feature selection technique. On the test dataset of 1,158 ECG recordings, we achieved an accuracy, sensitivity, and specificity of 95.3%, 77.7%, and 97.9%, respectively. Results demonstrate that our proposed hybrid method can extract features that are significant to detect AFib rhythms using single lead short ECG recordings.

Original languageEnglish (US)
Title of host publication2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages116-119
Number of pages4
ISBN (Electronic)9781538624050
DOIs
StatePublished - Apr 6 2018
Event2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018 - Las Vegas, United States
Duration: Mar 4 2018Mar 7 2018

Publication series

Name2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
Volume2018-January

Other

Other2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018
CountryUnited States
CityLas Vegas
Period3/4/183/7/18

Fingerprint

Electrocardiography
Atrial Fibrillation
Feature extraction
Entropy
Lead
Pattern recognition
Probability Theory
Cardiology
Classifiers
Genetic algorithms
Sensitivity and Specificity
Datasets

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Biomedical Engineering
  • Health Informatics

Cite this

Mahajan, R., Kamaleswaran, R., & Akbilgic, O. (2018). A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording. In 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018 (pp. 116-119). (2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018; Vol. 2018-January). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BHI.2018.8333383

A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording. / Mahajan, Ruhi; Kamaleswaran, Rishikesan; Akbilgic, Oguz.

2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 116-119 (2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018; Vol. 2018-January).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Mahajan, R, Kamaleswaran, R & Akbilgic, O 2018, A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording. in 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018. 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018, vol. 2018-January, Institute of Electrical and Electronics Engineers Inc., pp. 116-119, 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018, Las Vegas, United States, 3/4/18. https://doi.org/10.1109/BHI.2018.8333383
Mahajan R, Kamaleswaran R, Akbilgic O. A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording. In 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018. Institute of Electrical and Electronics Engineers Inc. 2018. p. 116-119. (2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018). https://doi.org/10.1109/BHI.2018.8333383
Mahajan, Ruhi ; Kamaleswaran, Rishikesan ; Akbilgic, Oguz. / A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording. 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018. Institute of Electrical and Electronics Engineers Inc., 2018. pp. 116-119 (2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018).
@inproceedings{af8ce6d89674458bb7a875d95b98b779,
title = "A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording",
abstract = "Identifying patients with Atrial Fibrillation (AFib) is one of the most challenging and prevailing problems in cardiology. In this study, we propose a novel feature extraction method hybridizing probabilistic symbolic pattern recognition (PSPR) and Sample Entropy (SampEn) to represent morphological changes in electrocardiogram (ECG) recordings. We implement a PSPR framework on continuous SampEn and RR interval series obtained from 4,630 ECG recordings in the training dataset. In our hybrid feature extraction method, PSPR symbolically discretizes SampEn and RR interval series with seven and nine unique symbols, respectively and then models the pattern transition behavior of these series using probability theory. We extract 28 features including PSPR-based metrics and descriptive metrics from SampEn, RR intervals, and processed ECG recordings. A random-forest classifier was trained on 13 features derived using a Genetic Algorithm based feature selection technique. On the test dataset of 1,158 ECG recordings, we achieved an accuracy, sensitivity, and specificity of 95.3{\%}, 77.7{\%}, and 97.9{\%}, respectively. Results demonstrate that our proposed hybrid method can extract features that are significant to detect AFib rhythms using single lead short ECG recordings.",
author = "Ruhi Mahajan and Rishikesan Kamaleswaran and Oguz Akbilgic",
year = "2018",
month = "4",
day = "6",
doi = "10.1109/BHI.2018.8333383",
language = "English (US)",
series = "2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "116--119",
booktitle = "2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018",
address = "United States",

}

TY - GEN

T1 - A hybrid feature extraction method to detect Atrial Fibrillation from single lead ECG recording

AU - Mahajan, Ruhi

AU - Kamaleswaran, Rishikesan

AU - Akbilgic, Oguz

PY - 2018/4/6

Y1 - 2018/4/6

N2 - Identifying patients with Atrial Fibrillation (AFib) is one of the most challenging and prevailing problems in cardiology. In this study, we propose a novel feature extraction method hybridizing probabilistic symbolic pattern recognition (PSPR) and Sample Entropy (SampEn) to represent morphological changes in electrocardiogram (ECG) recordings. We implement a PSPR framework on continuous SampEn and RR interval series obtained from 4,630 ECG recordings in the training dataset. In our hybrid feature extraction method, PSPR symbolically discretizes SampEn and RR interval series with seven and nine unique symbols, respectively and then models the pattern transition behavior of these series using probability theory. We extract 28 features including PSPR-based metrics and descriptive metrics from SampEn, RR intervals, and processed ECG recordings. A random-forest classifier was trained on 13 features derived using a Genetic Algorithm based feature selection technique. On the test dataset of 1,158 ECG recordings, we achieved an accuracy, sensitivity, and specificity of 95.3%, 77.7%, and 97.9%, respectively. Results demonstrate that our proposed hybrid method can extract features that are significant to detect AFib rhythms using single lead short ECG recordings.

AB - Identifying patients with Atrial Fibrillation (AFib) is one of the most challenging and prevailing problems in cardiology. In this study, we propose a novel feature extraction method hybridizing probabilistic symbolic pattern recognition (PSPR) and Sample Entropy (SampEn) to represent morphological changes in electrocardiogram (ECG) recordings. We implement a PSPR framework on continuous SampEn and RR interval series obtained from 4,630 ECG recordings in the training dataset. In our hybrid feature extraction method, PSPR symbolically discretizes SampEn and RR interval series with seven and nine unique symbols, respectively and then models the pattern transition behavior of these series using probability theory. We extract 28 features including PSPR-based metrics and descriptive metrics from SampEn, RR intervals, and processed ECG recordings. A random-forest classifier was trained on 13 features derived using a Genetic Algorithm based feature selection technique. On the test dataset of 1,158 ECG recordings, we achieved an accuracy, sensitivity, and specificity of 95.3%, 77.7%, and 97.9%, respectively. Results demonstrate that our proposed hybrid method can extract features that are significant to detect AFib rhythms using single lead short ECG recordings.

UR - http://www.scopus.com/inward/record.url?scp=85050866881&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85050866881&partnerID=8YFLogxK

U2 - 10.1109/BHI.2018.8333383

DO - 10.1109/BHI.2018.8333383

M3 - Conference contribution

AN - SCOPUS:85050866881

T3 - 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018

SP - 116

EP - 119

BT - 2018 IEEE EMBS International Conference on Biomedical and Health Informatics, BHI 2018

PB - Institute of Electrical and Electronics Engineers Inc.

ER -