Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration

Dunford, Carolyn E.; Marks, Nikki J.; Wilson, Rory; Scantlebury, D. Michael

doi:10.1002/ece3.11380

Journal article 21 views 6 downloads

Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration

Carolyn E. Dunford

, Nikki J. Marks, Rory Wilson

, D. Michael Scantlebury

Ecology and Evolution, Volume: 14, Issue: 5

Swansea University Author: Rory Wilson

PDF | Version of Record

© 2024 The Authors. This is an open access article under the terms of the Creative Commons Attribution License.
Download (1.68MB)

Check full text

DOI (Published version): 10.1002/ece3.11380

Abstract

Observing animals in the wild often poses extreme challenges, but animal-borne accelerometers are increasingly revealing unobservable behaviours. Automated machine learning streamlines behaviour identification from the substantial datasets generated during multi-animal, long-term studies; however, t...

Full description

Published in:	Ecology and Evolution
ISSN:	2045-7758 2045-7758
Published:	Wiley 2024
Online Access:	Check full text
URI:	https://cronfa.swan.ac.uk/Record/cronfa66202
Tags:	Add Tag No Tags, Be the first to tag this record!

first_indexed	2024-06-26T14:23:09Z
last_indexed	2024-06-26T14:23:09Z
id	cronfa66202
recordtype	SURis
fullrecord	<?xml version="1.0" encoding="utf-8"?><rfc1807 xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema"><bib-version>v2</bib-version><id>66202</id><entry>2024-04-25</entry><title>Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration</title><swanseaauthors><author><sid>017bc6dd155098860945dc6249c4e9bc</sid><ORCID>0000-0003-3177-0177</ORCID><firstname>Rory</firstname><surname>Wilson</surname><name>Rory Wilson</name><active>true</active><ethesisStudent>false</ethesisStudent></author></swanseaauthors><date>2024-04-25</date><deptcode>BGPS</deptcode><abstract>Observing animals in the wild often poses extreme challenges, but animal-borne accelerometers are increasingly revealing unobservable behaviours. Automated machine learning streamlines behaviour identification from the substantial datasets generated during multi-animal, long-term studies; however, the accuracy of such models depends on the qualities of the training data. We examined how data processing influenced the predictive accuracy of random forest (RF) models, leveraging the easily observed domestic cat (Felis catus) as a model organism for terrestrial mammalian behaviours. Nine indoor domestic cats were equipped with collar-mounted tri-axial accelerometers, and behaviours were recorded alongside video footage. From this calibrated data, eight datasets were derived with (i) additional descriptive variables, (ii) altered frequencies of acceleration data (40 Hz vs. a mean over 1 s) and (iii) standardised durations of different behaviours. These training datasets were used to generate RF models that were validated against calibrated cat behaviours before identifying the behaviours of five free-ranging tag-equipped cats. These predictions were compared to those identified manually to validate the accuracy of the RF models for free-ranging animal behaviours. RF models accurately predicted the behaviours of indoor domestic cats (F-measure up to 0.96) with discernible improvements observed with post-data-collection processing. Additional variables, standardised durations of behaviours and higher recording frequencies improved model accuracy. However, prediction accuracy varied with different behaviours, where high-frequency models excelled in identifying fast-paced behaviours (e.g. locomotion), whereas lower-frequency models (1 Hz) more accurately identified slower, aperiodic behaviours such as grooming and feeding, particularly when examining free-ranging cat behaviours. While RF modelling offered a robust means of behaviour identification from accelerometer data, field validations were important to validate model accuracy for free-ranging individuals. Future studies may benefit from employing similar data processing methods that enhance RF behaviour identification accuracy, with extensive advantages for investigations into ecology, welfare and management of wild animals.</abstract><type>Journal Article</type><journal>Ecology and Evolution</journal><volume>14</volume><journalNumber>5</journalNumber><paginationStart/><paginationEnd/><publisher>Wiley</publisher><placeOfPublication/><isbnPrint/><isbnElectronic/><issnPrint>2045-7758</issnPrint><issnElectronic>2045-7758</issnElectronic><keywords>activity budget, biologging, domestic cat, machine learning, predictive accuracy, random forestmodel</keywords><publishedDay>16</publishedDay><publishedMonth>5</publishedMonth><publishedYear>2024</publishedYear><publishedDate>2024-05-16</publishedDate><doi>10.1002/ece3.11380</doi><url/><notes/><college>COLLEGE NANME</college><department>Biosciences Geography and Physics School</department><CollegeCode>COLLEGE CODE</CollegeCode><DepartmentCode>BGPS</DepartmentCode><institution>Swansea University</institution><apcterm>Another institution paid the OA fee</apcterm><funders>Department for Education, UK Government</funders><projectreference/><lastEdited>2024-06-26T15:25:09.3506971</lastEdited><Created>2024-04-25T19:38:04.6803854</Created><path><level id="1">Faculty of Science and Engineering</level><level id="2">School of Biosciences, Geography and Physics - Biosciences</level></path><authors><author><firstname>Carolyn E.</firstname><surname>Dunford</surname><orcid>0000-0001-9850-4212</orcid><order>1</order></author><author><firstname>Nikki J.</firstname><surname>Marks</surname><order>2</order></author><author><firstname>Rory</firstname><surname>Wilson</surname><orcid>0000-0003-3177-0177</orcid><order>3</order></author><author><firstname>D. Michael</firstname><surname>Scantlebury</surname><orcid>0000-0001-8327-0556</orcid><order>4</order></author></authors><documents><document><filename>66202__30760__e11c74e2cef04168961c573365846178.pdf</filename><originalFilename>66202.VoR.pdf</originalFilename><uploaded>2024-06-26T15:23:54.9188234</uploaded><type>Output</type><contentLength>1758801</contentLength><contentType>application/pdf</contentType><version>Version of Record</version><cronfaStatus>true</cronfaStatus><documentNotes>© 2024 The Authors. This is an open access article under the terms of the Creative Commons Attribution License.</documentNotes><copyrightCorrect>true</copyrightCorrect><language>eng</language><licence>http://creativecommons.org/licenses/by/4.0/</licence></document></documents><OutputDurs/></rfc1807>
spelling	v2 66202 2024-04-25 Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration 017bc6dd155098860945dc6249c4e9bc 0000-0003-3177-0177 Rory Wilson Rory Wilson true false 2024-04-25 BGPS Observing animals in the wild often poses extreme challenges, but animal-borne accelerometers are increasingly revealing unobservable behaviours. Automated machine learning streamlines behaviour identification from the substantial datasets generated during multi-animal, long-term studies; however, the accuracy of such models depends on the qualities of the training data. We examined how data processing influenced the predictive accuracy of random forest (RF) models, leveraging the easily observed domestic cat (Felis catus) as a model organism for terrestrial mammalian behaviours. Nine indoor domestic cats were equipped with collar-mounted tri-axial accelerometers, and behaviours were recorded alongside video footage. From this calibrated data, eight datasets were derived with (i) additional descriptive variables, (ii) altered frequencies of acceleration data (40 Hz vs. a mean over 1 s) and (iii) standardised durations of different behaviours. These training datasets were used to generate RF models that were validated against calibrated cat behaviours before identifying the behaviours of five free-ranging tag-equipped cats. These predictions were compared to those identified manually to validate the accuracy of the RF models for free-ranging animal behaviours. RF models accurately predicted the behaviours of indoor domestic cats (F-measure up to 0.96) with discernible improvements observed with post-data-collection processing. Additional variables, standardised durations of behaviours and higher recording frequencies improved model accuracy. However, prediction accuracy varied with different behaviours, where high-frequency models excelled in identifying fast-paced behaviours (e.g. locomotion), whereas lower-frequency models (1 Hz) more accurately identified slower, aperiodic behaviours such as grooming and feeding, particularly when examining free-ranging cat behaviours. While RF modelling offered a robust means of behaviour identification from accelerometer data, field validations were important to validate model accuracy for free-ranging individuals. Future studies may benefit from employing similar data processing methods that enhance RF behaviour identification accuracy, with extensive advantages for investigations into ecology, welfare and management of wild animals. Journal Article Ecology and Evolution 14 5 Wiley 2045-7758 2045-7758 activity budget, biologging, domestic cat, machine learning, predictive accuracy, random forestmodel 16 5 2024 2024-05-16 10.1002/ece3.11380 COLLEGE NANME Biosciences Geography and Physics School COLLEGE CODE BGPS Swansea University Another institution paid the OA fee Department for Education, UK Government 2024-06-26T15:25:09.3506971 2024-04-25T19:38:04.6803854 Faculty of Science and Engineering School of Biosciences, Geography and Physics - Biosciences Carolyn E. Dunford 0000-0001-9850-4212 1 Nikki J. Marks 2 Rory Wilson 0000-0003-3177-0177 3 D. Michael Scantlebury 0000-0001-8327-0556 4 66202__30760__e11c74e2cef04168961c573365846178.pdf 66202.VoR.pdf 2024-06-26T15:23:54.9188234 Output 1758801 application/pdf Version of Record true © 2024 The Authors. This is an open access article under the terms of the Creative Commons Attribution License. true eng http://creativecommons.org/licenses/by/4.0/
title	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
spellingShingle	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration Rory Wilson
title_short	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
title_full	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
title_fullStr	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
title_full_unstemmed	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
title_sort	Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration
author_id_str_mv	017bc6dd155098860945dc6249c4e9bc
author_id_fullname_str_mv	017bc6dd155098860945dc6249c4e9bc_***_Rory Wilson
author	Rory Wilson
author2	Carolyn E. Dunford Nikki J. Marks Rory Wilson D. Michael Scantlebury
format	Journal article
container_title	Ecology and Evolution
container_volume	14
container_issue	5
publishDate	2024
institution	Swansea University
issn	2045-7758 2045-7758
doi_str_mv	10.1002/ece3.11380
publisher	Wiley
college_str	Faculty of Science and Engineering
hierarchytype
hierarchy_top_id	facultyofscienceandengineering
hierarchy_top_title	Faculty of Science and Engineering
hierarchy_parent_id	facultyofscienceandengineering
hierarchy_parent_title	Faculty of Science and Engineering
department_str	School of Biosciences, Geography and Physics - Biosciences{{{_:::_}}}Faculty of Science and Engineering{{{_:::_}}}School of Biosciences, Geography and Physics - Biosciences
document_store_str	1
active_str	0
description	Observing animals in the wild often poses extreme challenges, but animal-borne accelerometers are increasingly revealing unobservable behaviours. Automated machine learning streamlines behaviour identification from the substantial datasets generated during multi-animal, long-term studies; however, the accuracy of such models depends on the qualities of the training data. We examined how data processing influenced the predictive accuracy of random forest (RF) models, leveraging the easily observed domestic cat (Felis catus) as a model organism for terrestrial mammalian behaviours. Nine indoor domestic cats were equipped with collar-mounted tri-axial accelerometers, and behaviours were recorded alongside video footage. From this calibrated data, eight datasets were derived with (i) additional descriptive variables, (ii) altered frequencies of acceleration data (40 Hz vs. a mean over 1 s) and (iii) standardised durations of different behaviours. These training datasets were used to generate RF models that were validated against calibrated cat behaviours before identifying the behaviours of five free-ranging tag-equipped cats. These predictions were compared to those identified manually to validate the accuracy of the RF models for free-ranging animal behaviours. RF models accurately predicted the behaviours of indoor domestic cats (F-measure up to 0.96) with discernible improvements observed with post-data-collection processing. Additional variables, standardised durations of behaviours and higher recording frequencies improved model accuracy. However, prediction accuracy varied with different behaviours, where high-frequency models excelled in identifying fast-paced behaviours (e.g. locomotion), whereas lower-frequency models (1 Hz) more accurately identified slower, aperiodic behaviours such as grooming and feeding, particularly when examining free-ranging cat behaviours. While RF modelling offered a robust means of behaviour identification from accelerometer data, field validations were important to validate model accuracy for free-ranging individuals. Future studies may benefit from employing similar data processing methods that enhance RF behaviour identification accuracy, with extensive advantages for investigations into ecology, welfare and management of wild animals.
published_date	2024-05-16T15:25:08Z
_version_	1802934061168066560
score	11.012678

Identifying animal behaviours from accelerometers: Improving predictive accuracy of machine learning by refining the variables selected, data frequency, and sample duration

Similar Items