Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience

Hardik Raipal; Madalina Sas; Chris Lockwood; Rebecca Joakim; Nicholas S. Peters; Max Falkenberg

doi:10.22489/CinC.2020.185

Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience

Hardik Raipal, Madalina Sas, Chris Lockwood, Rebecca Joakim, Nicholas S. Peters, Max Falkenberg

Research output: Contribution to Book/Report types › Conference contribution › peer-review

Abstract (may include machine translation)

Automated ECG classification is a standard feature in many commercial 12-Lead ECG machines. As part of the Physionet/CinC Challenge 2020, our team, 'Mad-hardmax', developed an XGBoost based classification method for the analysis of 12-Lead ECGs acquired from four different countries. Our aim is to develop an interpretable classifier that outputs diagnoses which can be traced to specific ECG features, while also testing the potential of information theoretic features for ECG diagnosis. These measures capture high-level interdependencies across ECG leads which are effective for discriminating conditions with multiple complex morphologies. On unseen test data, our algorithm achieved a challenge score of 0.155 relative to a winning score of 0.533, putting our submission in 24th position from 41 successful entries.

Original language	English
Title of host publication	2020 Computing in Cardiology, CinC 2020
Publisher	IEEE Computer Society
ISBN (Electronic)	9781728173825
DOIs	https://doi.org/10.22489/CinC.2020.185
State	Published - 13 Sep 2020
Externally published	Yes
Event	2020 Computing in Cardiology, CinC 2020 - Rimini, Italy Duration: 13 Sep 2020 → 16 Sep 2020

Publication series

Name	Computing in Cardiology
Volume	2020-September
ISSN (Print)	2325-8861
ISSN (Electronic)	2325-887X

Conference

Conference	2020 Computing in Cardiology, CinC 2020
Country/Territory	Italy
City	Rimini
Period	13/09/20 → 16/09/20

Access to Document

10.22489/CinC.2020.185License: Unspecified

Cite this

Raipal, H., Sas, M., Lockwood, C., Joakim, R., Peters, N. S., & Falkenberg, M. (2020). Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience. In 2020 Computing in Cardiology, CinC 2020 Article 9344417 (Computing in Cardiology; Vol. 2020-September). IEEE Computer Society. https://doi.org/10.22489/CinC.2020.185

@inproceedings{9c9415685d6c40808eed644fffa65379,

title = "Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience",

abstract = "Automated ECG classification is a standard feature in many commercial 12-Lead ECG machines. As part of the Physionet/CinC Challenge 2020, our team, 'Mad-hardmax', developed an XGBoost based classification method for the analysis of 12-Lead ECGs acquired from four different countries. Our aim is to develop an interpretable classifier that outputs diagnoses which can be traced to specific ECG features, while also testing the potential of information theoretic features for ECG diagnosis. These measures capture high-level interdependencies across ECG leads which are effective for discriminating conditions with multiple complex morphologies. On unseen test data, our algorithm achieved a challenge score of 0.155 relative to a winning score of 0.533, putting our submission in 24th position from 41 successful entries.",

author = "Hardik Raipal and Madalina Sas and Chris Lockwood and Rebecca Joakim and Peters, {Nicholas S.} and Max Falkenberg",

note = "Publisher Copyright: {\textcopyright} 2020 Creative Commons; the authors hold their copyright.; 2020 Computing in Cardiology, CinC 2020 ; Conference date: 13-09-2020 Through 16-09-2020",

year = "2020",

month = sep,

day = "13",

doi = "10.22489/CinC.2020.185",

language = "English",

series = "Computing in Cardiology",

publisher = "IEEE Computer Society",

booktitle = "2020 Computing in Cardiology, CinC 2020",

}

Raipal, H, Sas, M, Lockwood, C, Joakim, R, Peters, NS & Falkenberg, M 2020, Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience. in 2020 Computing in Cardiology, CinC 2020., 9344417, Computing in Cardiology, vol. 2020-September, IEEE Computer Society, 2020 Computing in Cardiology, CinC 2020, Rimini, Italy, 13/09/20. https://doi.org/10.22489/CinC.2020.185

Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience. / Raipal, Hardik; Sas, Madalina; Lockwood, Chris et al.
2020 Computing in Cardiology, CinC 2020. IEEE Computer Society, 2020. 9344417 (Computing in Cardiology; Vol. 2020-September).

Research output: Contribution to Book/Report types › Conference contribution › peer-review

TY - GEN

T1 - Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience

AU - Raipal, Hardik

AU - Sas, Madalina

AU - Lockwood, Chris

AU - Joakim, Rebecca

AU - Peters, Nicholas S.

AU - Falkenberg, Max

PY - 2020/9/13

Y1 - 2020/9/13

N2 - Automated ECG classification is a standard feature in many commercial 12-Lead ECG machines. As part of the Physionet/CinC Challenge 2020, our team, 'Mad-hardmax', developed an XGBoost based classification method for the analysis of 12-Lead ECGs acquired from four different countries. Our aim is to develop an interpretable classifier that outputs diagnoses which can be traced to specific ECG features, while also testing the potential of information theoretic features for ECG diagnosis. These measures capture high-level interdependencies across ECG leads which are effective for discriminating conditions with multiple complex morphologies. On unseen test data, our algorithm achieved a challenge score of 0.155 relative to a winning score of 0.533, putting our submission in 24th position from 41 successful entries.

AB - Automated ECG classification is a standard feature in many commercial 12-Lead ECG machines. As part of the Physionet/CinC Challenge 2020, our team, 'Mad-hardmax', developed an XGBoost based classification method for the analysis of 12-Lead ECGs acquired from four different countries. Our aim is to develop an interpretable classifier that outputs diagnoses which can be traced to specific ECG features, while also testing the potential of information theoretic features for ECG diagnosis. These measures capture high-level interdependencies across ECG leads which are effective for discriminating conditions with multiple complex morphologies. On unseen test data, our algorithm achieved a challenge score of 0.155 relative to a winning score of 0.533, putting our submission in 24th position from 41 successful entries.

UR - http://www.scopus.com/inward/record.url?scp=85100946509&partnerID=8YFLogxK

U2 - 10.22489/CinC.2020.185

DO - 10.22489/CinC.2020.185

M3 - Conference contribution

AN - SCOPUS:85100946509

T3 - Computing in Cardiology

BT - 2020 Computing in Cardiology, CinC 2020

PB - IEEE Computer Society

T2 - 2020 Computing in Cardiology, CinC 2020

Y2 - 13 September 2020 through 16 September 2020

ER -

Interpretable XGBoost Based Classification of 12-lead ECGs Applying Information Theory Measures from Neuroscience

Abstract (may include machine translation)

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this