{"created":"2025-01-19T01:01:14.047570+00:00","metadata":{"_oai":{"id":"oai:ipsj.ixsq.nii.ac.jp:00196717","sets":["1164:4061:9694:9777"]},"path":["9777"],"owner":"44499","recid":"196717","title":["Predicting Medical-Grade Sleep-Wake Classification from Fitbit Data Using Tree-Based Machine Learning"],"pubdate":{"attribute_name":"公開日","attribute_value":"2019-05-30"},"_buckets":{"deposit":"76b82f7c-fec8-4146-9376-f244ca0c317f"},"_deposit":{"id":"196717","pid":{"type":"depid","value":"196717","revision_id":0},"owners":[44499],"status":"published","created_by":44499},"item_title":"Predicting Medical-Grade Sleep-Wake Classification from Fitbit Data Using Tree-Based Machine Learning","author_link":["471088","471090","471091","471089"],"item_titles":{"attribute_name":"タイトル","attribute_value_mlt":[{"subitem_title":"Predicting Medical-Grade Sleep-Wake Classification from Fitbit Data Using Tree-Based Machine Learning"},{"subitem_title":"Predicting Medical-Grade Sleep-Wake Classification from Fitbit Data Using Tree-Based Machine Learning","subitem_title_language":"en"}]},"item_keyword":{"attribute_name":"キーワード","attribute_value_mlt":[{"subitem_subject":"行動認識","subitem_subject_scheme":"Other"}]},"item_type_id":"4","publish_date":"2019-05-30","item_4_text_3":{"attribute_name":"著者所属","attribute_value_mlt":[{"subitem_text_value":"Kyoto University of Advanced Science／The University of Tokyo"},{"subitem_text_value":"CAC Corporation"}]},"item_4_text_4":{"attribute_name":"著者所属(英)","attribute_value_mlt":[{"subitem_text_value":"Kyoto University of Advanced Science / The University of Tokyo","subitem_text_language":"en"},{"subitem_text_value":"CAC Corporation","subitem_text_language":"en"}]},"item_language":{"attribute_name":"言語","attribute_value_mlt":[{"subitem_language":"eng"}]},"item_publisher":{"attribute_name":"出版者","attribute_value_mlt":[{"subitem_publisher":"情報処理学会","subitem_publisher_language":"ja"}]},"publish_status":"0","weko_shared_id":-1,"item_file_price":{"attribute_name":"Billing file","attribute_type":"file","attribute_value_mlt":[{"url":{"url":"https://ipsj.ixsq.nii.ac.jp/record/196717/files/IPSJ-UBI19062014.pdf","label":"IPSJ-UBI19062014.pdf"},"date":[{"dateType":"Available","dateValue":"2021-05-30"}],"format":"application/pdf","billing":["billing_file"],"filename":"IPSJ-UBI19062014.pdf","filesize":[{"value":"1.9 MB"}],"mimetype":"application/pdf","priceinfo":[{"tax":["include_tax"],"price":"660","billingrole":"5"},{"tax":["include_tax"],"price":"330","billingrole":"6"},{"tax":["include_tax"],"price":"0","billingrole":"36"},{"tax":["include_tax"],"price":"0","billingrole":"44"}],"accessrole":"open_date","version_id":"d6989afa-5309-4e27-9c11-1bd5a3eb5096","displaytype":"detail","licensetype":"license_note","license_note":"Copyright (c) 2019 by the Information Processing Society of Japan"}]},"item_4_creator_5":{"attribute_name":"著者名","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Zilu, Liang"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Mario, Alberto Chapa-Martell"}],"nameIdentifiers":[{}]}]},"item_4_creator_6":{"attribute_name":"著者名(英)","attribute_type":"creator","attribute_value_mlt":[{"creatorNames":[{"creatorName":"Zilu, Liang","creatorNameLang":"en"}],"nameIdentifiers":[{}]},{"creatorNames":[{"creatorName":"Mario, Alberto Chapa-Martell","creatorNameLang":"en"}],"nameIdentifiers":[{}]}]},"item_4_source_id_9":{"attribute_name":"書誌レコードID","attribute_value_mlt":[{"subitem_source_identifier":"AA11838947","subitem_source_identifier_type":"NCID"}]},"item_4_textarea_12":{"attribute_name":"Notice","attribute_value_mlt":[{"subitem_textarea_value":"SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc."}]},"item_resource_type":{"attribute_name":"資源タイプ","attribute_value_mlt":[{"resourceuri":"http://purl.org/coar/resource_type/c_18gh","resourcetype":"technical report"}]},"item_4_source_id_11":{"attribute_name":"ISSN","attribute_value_mlt":[{"subitem_source_identifier":"2188-8698","subitem_source_identifier_type":"ISSN"}]},"item_4_description_7":{"attribute_name":"論文抄録","attribute_value_mlt":[{"subitem_description":"It has become increasingly popular among individuals and researchers to monitor sleep using consumer activity wristbands. Nevertheless, many validation studies have identified a significant gap between consumer wristbands and medical sleep monitors. This study aims to bridge this gap through developing predictive models that leverage Fitbit data to generate medical-grade sleep/wake classification. Considering that the “sleep” class significantly outnumbers the “wake” class, we formulated the problem of interest into an imbalanced classification problem. We applied two tree-based machine learning techniques, i.e. decision tree and random forest, in combination with four re-sampling methods, which yields in total 10 classifiers. The performance of the classifiers was compared to the original Fitbit algorithm based on sensitivity, specificity and area under the ROC curve (AUC). Our results showed that in the best case, specificity was improved by 75% while sensitivity was reduced by 12%, which yielded a statistically significant increase of 11% in AUC. The decision tree technique was more robust and less affected by re-sampling method compared to the random forest technique, and random up sampling may be a most effective re-sampling strategy to balance the training sets. These findings demonstrate the feasibility of achieving medical-grade sleep/wake classification from consumer wristbands by applying proper combination of reesampling and machine techniques.","subitem_description_type":"Other"}]},"item_4_description_8":{"attribute_name":"論文抄録(英)","attribute_value_mlt":[{"subitem_description":"It has become increasingly popular among individuals and researchers to monitor sleep using consumer activity wristbands. Nevertheless, many validation studies have identified a significant gap between consumer wristbands and medical sleep monitors. This study aims to bridge this gap through developing predictive models that leverage Fitbit data to generate medical-grade sleep/wake classification. Considering that the “sleep” class significantly outnumbers the “wake” class, we formulated the problem of interest into an imbalanced classification problem. We applied two tree-based machine learning techniques, i.e. decision tree and random forest, in combination with four re-sampling methods, which yields in total 10 classifiers. The performance of the classifiers was compared to the original Fitbit algorithm based on sensitivity, specificity and area under the ROC curve (AUC). Our results showed that in the best case, specificity was improved by 75% while sensitivity was reduced by 12%, which yielded a statistically significant increase of 11% in AUC. The decision tree technique was more robust and less affected by re-sampling method compared to the random forest technique, and random up sampling may be a most effective re-sampling strategy to balance the training sets. These findings demonstrate the feasibility of achieving medical-grade sleep/wake classification from consumer wristbands by applying proper combination of reesampling and machine techniques.","subitem_description_type":"Other"}]},"item_4_biblio_info_10":{"attribute_name":"書誌情報","attribute_value_mlt":[{"bibliographicPageEnd":"8","bibliographic_titles":[{"bibliographic_title":"研究報告ユビキタスコンピューティングシステム（UBI）"}],"bibliographicPageStart":"1","bibliographicIssueDates":{"bibliographicIssueDate":"2019-05-30","bibliographicIssueDateType":"Issued"},"bibliographicIssueNumber":"14","bibliographicVolumeNumber":"2019-UBI-62"}]},"relation_version_is_last":true,"weko_creator_id":"44499"},"links":{},"id":196717,"updated":"2025-01-19T22:39:58.823293+00:00"}