No Cover Image

Journal article 907 views 418 downloads

Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth

Alexander H.K. Montoye, Kimberly A. Clevenger, Kelly Mackintosh Orcid Logo, Melitta McNarry Orcid Logo, Karin A. Pfeiffer

Journal for the Measurement of Physical Behaviour, Volume: 2, Issue: 4, Pages: 237 - 246

Swansea University Authors: Kelly Mackintosh Orcid Logo, Melitta McNarry Orcid Logo

Check full text

DOI (Published version): 10.1123/jmpb.2018-0011

Abstract

Background: Machine learning may improve energy expenditure (EE) prediction from body-worn accelerometers. However, machine learning models are rarely cross-validated in an independent sample, and the use of machine learning raises additional questions including the effect of accelerometer placement...

Full description

Published in: Journal for the Measurement of Physical Behaviour
ISSN: 2575-6605 2575-6613
Published: Human Kinetics 2019
Online Access: Check full text

URI: https://cronfa.swan.ac.uk/Record/cronfa50558
Tags: Add Tag
No Tags, Be the first to tag this record!
first_indexed 2019-06-05T11:07:49Z
last_indexed 2020-06-26T19:01:45Z
id cronfa50558
recordtype SURis
fullrecord <?xml version="1.0"?><rfc1807><datestamp>2020-06-26T17:56:32.7090435</datestamp><bib-version>v2</bib-version><id>50558</id><entry>2019-05-29</entry><title>Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth</title><swanseaauthors><author><sid>bdb20e3f31bcccf95c7bc116070c4214</sid><ORCID>0000-0003-0355-6357</ORCID><firstname>Kelly</firstname><surname>Mackintosh</surname><name>Kelly Mackintosh</name><active>true</active><ethesisStudent>false</ethesisStudent></author><author><sid>062f5697ff59f004bc8c713955988398</sid><ORCID>0000-0003-0813-7477</ORCID><firstname>Melitta</firstname><surname>McNarry</surname><name>Melitta McNarry</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2019-05-29</date><deptcode>STSC</deptcode><abstract>Background: Machine learning may improve energy expenditure (EE) prediction from body-worn accelerometers. However, machine learning models are rarely cross-validated in an independent sample, and the use of machine learning raises additional questions including the effect of accelerometer placement and data type (count vs. raw) for optimal EE prediction. Purpose: To assess the accuracy of artificial neural network (ANN) models for EE prediction in youth using count-based or raw data from accelerometers worn on the hip, wrist, or in combination, and compare these to count-based, EE regression equations. Methods: Data were collected in two settings; one (n&#x2009;=&#x2009;27) to calibrate the EE prediction models, and the other (n&#x2009;=&#x2009;34) for model cross-validation. Participants wore a portable metabolic analyzer (EE criterion) and accelerometers on the left wrist and right hip while completing 30 minutes of exergames (calibration, cross-validation) and a maximal exercise test (calibration only). Six ANNs were created from the calibration data, separately by accelerometer placement (hip, wrist, combination) and data format (count-based, raw) to predict EE (15-second epochs). Three count-based linear regression equations were also developed for comparison to the ANNs. Results: The count-based, hip ANN demonstrated lower error (RMSE: 1.2 METs) than all other ANNs (RMSE: 1.7&#x2013;3.6 METs) and EE regression equations (RMSE: 1.5&#x2013;3.2 METs). However, all models showed bias toward the mean. Conclusion: An ANN developed for hip-worn accelerometers had higher accuracy for EE prediction during an exergame session than wrist or combination ANNs, and ANNs developed using count-based data had higher accuracy than ANNs developed using raw data.</abstract><type>Journal Article</type><journal>Journal for the Measurement of Physical Behaviour</journal><volume>2</volume><journalNumber>4</journalNumber><paginationStart>237</paginationStart><paginationEnd>246</paginationEnd><publisher>Human Kinetics</publisher><issnPrint>2575-6605</issnPrint><issnElectronic>2575-6613</issnElectronic><keywords/><publishedDay>31</publishedDay><publishedMonth>12</publishedMonth><publishedYear>2019</publishedYear><publishedDate>2019-12-31</publishedDate><doi>10.1123/jmpb.2018-0011</doi><url/><notes/><college>COLLEGE NANME</college><department>Sport and Exercise Sciences</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>STSC</DepartmentCode><institution>Swansea University</institution><apcterm/><lastEdited>2020-06-26T17:56:32.7090435</lastEdited><Created>2019-05-29T11:02:32.7032214</Created><authors><author><firstname>Alexander H.K.</firstname><surname>Montoye</surname><order>1</order></author><author><firstname>Kimberly A.</firstname><surname>Clevenger</surname><order>2</order></author><author><firstname>Kelly</firstname><surname>Mackintosh</surname><orcid>0000-0003-0355-6357</orcid><order>3</order></author><author><firstname>Melitta</firstname><surname>McNarry</surname><orcid>0000-0003-0813-7477</orcid><order>4</order></author><author><firstname>Karin A.</firstname><surname>Pfeiffer</surname><order>5</order></author></authors><documents><document><filename>50558__14047__d988df4757c14d7cb4c5c324be577406.pdf</filename><originalFilename>MontoyeFigure2.pdf</originalFilename><uploaded>2019-05-29T11:42:42.2470000</uploaded><type>Output</type><contentLength>1823189</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2020-12-01T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document><document><filename>50558__14046__800f28faf7e2401e881935aebe64da78.pdf</filename><originalFilename>MontoyeFigure1.pdf</originalFilename><uploaded>2019-05-29T11:41:48.8570000</uploaded><type>Output</type><contentLength>55022</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2020-12-01T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document><document><filename>50558__14044__b48d6bbcbfbd408f986dffe73ea4795c.pdf</filename><originalFilename>Montoye2019.pdf</originalFilename><uploaded>2019-05-29T11:40:31.0500000</uploaded><type>Output</type><contentLength>597842</contentLength><contentType>application/pdf</contentType><version>Accepted Manuscript</version><cronfaStatus>true</cronfaStatus><embargoDate>2019-12-01T00:00:00.0000000</embargoDate><copyrightCorrect>true</copyrightCorrect><language>eng</language></document></documents><OutputDurs/></rfc1807>
spelling 2020-06-26T17:56:32.7090435 v2 50558 2019-05-29 Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth bdb20e3f31bcccf95c7bc116070c4214 0000-0003-0355-6357 Kelly Mackintosh Kelly Mackintosh true false 062f5697ff59f004bc8c713955988398 0000-0003-0813-7477 Melitta McNarry Melitta McNarry true false 2019-05-29 STSC Background: Machine learning may improve energy expenditure (EE) prediction from body-worn accelerometers. However, machine learning models are rarely cross-validated in an independent sample, and the use of machine learning raises additional questions including the effect of accelerometer placement and data type (count vs. raw) for optimal EE prediction. Purpose: To assess the accuracy of artificial neural network (ANN) models for EE prediction in youth using count-based or raw data from accelerometers worn on the hip, wrist, or in combination, and compare these to count-based, EE regression equations. Methods: Data were collected in two settings; one (n = 27) to calibrate the EE prediction models, and the other (n = 34) for model cross-validation. Participants wore a portable metabolic analyzer (EE criterion) and accelerometers on the left wrist and right hip while completing 30 minutes of exergames (calibration, cross-validation) and a maximal exercise test (calibration only). Six ANNs were created from the calibration data, separately by accelerometer placement (hip, wrist, combination) and data format (count-based, raw) to predict EE (15-second epochs). Three count-based linear regression equations were also developed for comparison to the ANNs. Results: The count-based, hip ANN demonstrated lower error (RMSE: 1.2 METs) than all other ANNs (RMSE: 1.7–3.6 METs) and EE regression equations (RMSE: 1.5–3.2 METs). However, all models showed bias toward the mean. Conclusion: An ANN developed for hip-worn accelerometers had higher accuracy for EE prediction during an exergame session than wrist or combination ANNs, and ANNs developed using count-based data had higher accuracy than ANNs developed using raw data. Journal Article Journal for the Measurement of Physical Behaviour 2 4 237 246 Human Kinetics 2575-6605 2575-6613 31 12 2019 2019-12-31 10.1123/jmpb.2018-0011 COLLEGE NANME Sport and Exercise Sciences COLLEGE CODE STSC Swansea University 2020-06-26T17:56:32.7090435 2019-05-29T11:02:32.7032214 Alexander H.K. Montoye 1 Kimberly A. Clevenger 2 Kelly Mackintosh 0000-0003-0355-6357 3 Melitta McNarry 0000-0003-0813-7477 4 Karin A. Pfeiffer 5 50558__14047__d988df4757c14d7cb4c5c324be577406.pdf MontoyeFigure2.pdf 2019-05-29T11:42:42.2470000 Output 1823189 application/pdf Accepted Manuscript true 2020-12-01T00:00:00.0000000 true eng 50558__14046__800f28faf7e2401e881935aebe64da78.pdf MontoyeFigure1.pdf 2019-05-29T11:41:48.8570000 Output 55022 application/pdf Accepted Manuscript true 2020-12-01T00:00:00.0000000 true eng 50558__14044__b48d6bbcbfbd408f986dffe73ea4795c.pdf Montoye2019.pdf 2019-05-29T11:40:31.0500000 Output 597842 application/pdf Accepted Manuscript true 2019-12-01T00:00:00.0000000 true eng
title Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
spellingShingle Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
Kelly Mackintosh
Melitta McNarry
title_short Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
title_full Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
title_fullStr Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
title_full_unstemmed Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
title_sort Cross-Validation and Comparison of Energy Expenditure Prediction Models Using Count-Based and Raw Accelerometer Data in Youth
author_id_str_mv bdb20e3f31bcccf95c7bc116070c4214
062f5697ff59f004bc8c713955988398
author_id_fullname_str_mv bdb20e3f31bcccf95c7bc116070c4214_***_Kelly Mackintosh
062f5697ff59f004bc8c713955988398_***_Melitta McNarry
author Kelly Mackintosh
Melitta McNarry
author2 Alexander H.K. Montoye
Kimberly A. Clevenger
Kelly Mackintosh
Melitta McNarry
Karin A. Pfeiffer
format Journal article
container_title Journal for the Measurement of Physical Behaviour
container_volume 2
container_issue 4
container_start_page 237
publishDate 2019
institution Swansea University
issn 2575-6605
2575-6613
doi_str_mv 10.1123/jmpb.2018-0011
publisher Human Kinetics
document_store_str 1
active_str 0
description Background: Machine learning may improve energy expenditure (EE) prediction from body-worn accelerometers. However, machine learning models are rarely cross-validated in an independent sample, and the use of machine learning raises additional questions including the effect of accelerometer placement and data type (count vs. raw) for optimal EE prediction. Purpose: To assess the accuracy of artificial neural network (ANN) models for EE prediction in youth using count-based or raw data from accelerometers worn on the hip, wrist, or in combination, and compare these to count-based, EE regression equations. Methods: Data were collected in two settings; one (n = 27) to calibrate the EE prediction models, and the other (n = 34) for model cross-validation. Participants wore a portable metabolic analyzer (EE criterion) and accelerometers on the left wrist and right hip while completing 30 minutes of exergames (calibration, cross-validation) and a maximal exercise test (calibration only). Six ANNs were created from the calibration data, separately by accelerometer placement (hip, wrist, combination) and data format (count-based, raw) to predict EE (15-second epochs). Three count-based linear regression equations were also developed for comparison to the ANNs. Results: The count-based, hip ANN demonstrated lower error (RMSE: 1.2 METs) than all other ANNs (RMSE: 1.7–3.6 METs) and EE regression equations (RMSE: 1.5–3.2 METs). However, all models showed bias toward the mean. Conclusion: An ANN developed for hip-worn accelerometers had higher accuracy for EE prediction during an exergame session than wrist or combination ANNs, and ANNs developed using count-based data had higher accuracy than ANNs developed using raw data.
published_date 2019-12-31T04:02:03Z
_version_ 1763753195684757504
score 11.016235

Similar Items