{"id":774,"date":"2023-03-23T09:20:00","date_gmt":"2023-03-23T09:20:00","guid":{"rendered":"https:\/\/igormatias.com\/?p=774"},"modified":"2026-02-02T13:04:43","modified_gmt":"2026-02-02T13:04:43","slug":"deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast","status":"publish","type":"post","link":"https:\/\/igormatias.com\/pt\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/","title":{"rendered":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast"},"content":{"rendered":"

Abstract: <\/span>Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast to the traditional clinical environment, these devices allow time series data collection in an individual\u2019s daily living environ- ment. However, missing data can occur while using them. Many techniques have been applied to solve these data gaps; nonetheless, missing time series data poses extra challenges, such as maintaining the temporal dependency. In this article, we addressed the forecast of sleep trackers data (sleeping heart rate (HR) and time asleep) for 2 main reasons: (1) to design models capable of accurately forecasting missing data from those devices, and (2) to apply those models to empower sleep interventions that may increase its quality, by forecasting future sleep events. We collected wearables data over 290 days (per individual) from 12 participants using a smartwatch and made this dataset publicly available. We then explored several hyperparameters of 2 Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). We further elaborated and compared the performance of 3 approaches to training those RNNs. Although similar performance, slightly more accurate results were obtained after training a GRU network on an entire population\u2019s dataset, which was able to forecast the average, minimum, and maximum sleeping HR with a root-mean-squared error (RMSE) of 4.4 (\u00b1 1.4), 4.9 (\u00b1 2.6), and 12.1 (\u00b1 4.0) beats per minute, respectively. However, the total time asleep was impossible to forecast with low error.<\/p>\n

Igor Matias1<\/sup>, and Katarzyna Wac1<\/sup><\/p>\n

1<\/sup>Quality of Life Technologies Lab, University of Geneva, Switzerland<\/p>\n

In 2022 IEEE International Conference on Machine Learning and Applications (ICMLA)<\/span><\/p>\n

https:\/\/doi.org\/10.1109\/ICMLA55696.2022.00264<\/a><\/p>\n\n\n

<\/p>","protected":false},"excerpt":{"rendered":"

Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast to the traditional clinical environment, these devices allow time series data collection in an individual\u2019s daily living environ- ment. However, missing data can occur while using them. Many techniques have been applied to solve these data gaps; nonetheless, missing time series data poses extra challenges, such as maintaining the temporal dependency. In this article, we addressed the forecast of sleep trackers data (sleeping heart rate (HR) and time asleep) for 2 main reasons: (1) to design models capable of accurately forecasting missing data from those devices, and (2) to apply those models to empower sleep interventions that may increase its quality, by forecasting future sleep events. We collected wearables data over 290 days (per individual) from 12 participants using a smartwatch and made this dataset publicly available. We then explored several hyperparameters of 2 Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU). We further elaborated and compared the performance of 3 approaches to training those RNNs. Although similar performance, slightly more accurate results were obtained after training a GRU network on an entire population\u2019s dataset, which was able to forecast the average, minimum, and maximum sleeping HR with a root-mean-squared error (RMSE) of 4.4 (\u00b1 1.4), 4.9 (\u00b1 2.6), and 12.1 (\u00b1 4.0) beats per minute, respectively. However, the total time asleep was impossible to forecast with low error.<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":"","_links_to":"","_links_to_target":""},"categories":[11],"tags":[],"class_list":["post-774","post","type-post","status-publish","format-standard","hentry","category-conference"],"gutentor_comment":0,"yoast_head":"\nDej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast<\/title>\n<meta name=\"description\" content=\"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/igormatias.com\/pt\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\" \/>\n<meta property=\"og:locale\" content=\"pt_PT\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast\" \/>\n<meta property=\"og:description\" content=\"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...\" \/>\n<meta property=\"og:url\" content=\"https:\/\/igormatias.com\/pt\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\" \/>\n<meta property=\"og:site_name\" content=\"Igor Matias 🧠🤖\" \/>\n<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/igormatias16\" \/>\n<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/igormatias16\" \/>\n<meta property=\"article:published_time\" content=\"2023-03-23T09:20:00+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-02-02T13:04:43+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"933\" \/>\n\t<meta property=\"og:image:height\" content=\"933\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Igor Matias\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@IgorMatias3\" \/>\n<meta name=\"twitter:site\" content=\"@IgorMatias3\" \/>\n<meta name=\"twitter:label1\" content=\"Escrito por\" \/>\n\t<meta name=\"twitter:data1\" content=\"Igor Matias\" \/>\n\t<meta name=\"twitter:label2\" content=\"Tempo estimado de leitura\" \/>\n\t<meta name=\"twitter:data2\" content=\"2 minutos\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\"},\"author\":{\"name\":\"Igor Matias\",\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248\"},\"headline\":\"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast\",\"datePublished\":\"2023-03-23T09:20:00+00:00\",\"dateModified\":\"2026-02-02T13:04:43+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\"},\"wordCount\":266,\"publisher\":{\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248\"},\"articleSection\":[\"Conference\"],\"inLanguage\":\"pt-PT\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\",\"url\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\",\"name\":\"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast\",\"isPartOf\":{\"@id\":\"https:\/\/igormatias.com\/#website\"},\"datePublished\":\"2023-03-23T09:20:00+00:00\",\"dateModified\":\"2026-02-02T13:04:43+00:00\",\"description\":\"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...\",\"breadcrumb\":{\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#breadcrumb\"},\"inLanguage\":\"pt-PT\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/igormatias.com\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/igormatias.com\/#website\",\"url\":\"https:\/\/igormatias.com\/\",\"name\":\"Igor Matias\",\"description\":\"(Computer Science, Health, AI)\",\"publisher\":{\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/igormatias.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"pt-PT\"},{\"@type\":[\"Person\",\"Organization\"],\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248\",\"name\":\"Igor Matias\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"pt-PT\",\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg\",\"contentUrl\":\"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg\",\"width\":933,\"height\":933,\"caption\":\"Igor Matias\"},\"logo\":{\"@id\":\"https:\/\/igormatias.com\/#\/schema\/person\/image\/\"},\"description\":\"Ph.D. student in Information Systems at the University of Geneva, Switzerland, focusing Alzheimer\u2019s Disease diagnosis. Master of Computer Science and Engineering by University of Beira Interior, Portugal.\",\"sameAs\":[\"https:\/\/igormatias.com\/\",\"https:\/\/www.facebook.com\/igormatias16\",\"https:\/\/instagram.com\/igormatias3\",\"https:\/\/linkedin.com\/in\/igormatias3\",\"https:\/\/x.com\/IgorMatias3\"]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast","description":"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/igormatias.com\/pt\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/","og_locale":"pt_PT","og_type":"article","og_title":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast","og_description":"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...","og_url":"https:\/\/igormatias.com\/pt\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/","og_site_name":"Igor Matias 🧠🤖","article_publisher":"https:\/\/www.facebook.com\/igormatias16","article_author":"https:\/\/www.facebook.com\/igormatias16","article_published_time":"2023-03-23T09:20:00+00:00","article_modified_time":"2026-02-02T13:04:43+00:00","og_image":[{"width":933,"height":933,"url":"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg","type":"image\/jpeg"}],"author":"Igor Matias","twitter_card":"summary_large_image","twitter_creator":"@IgorMatias3","twitter_site":"@IgorMatias3","twitter_misc":{"Escrito por":"Igor Matias","Tempo estimado de leitura":"2 minutos"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#article","isPartOf":{"@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/"},"author":{"name":"Igor Matias","@id":"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248"},"headline":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast","datePublished":"2023-03-23T09:20:00+00:00","dateModified":"2026-02-02T13:04:43+00:00","mainEntityOfPage":{"@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/"},"wordCount":266,"publisher":{"@id":"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248"},"articleSection":["Conference"],"inLanguage":"pt-PT"},{"@type":"WebPage","@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/","url":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/","name":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast","isPartOf":{"@id":"https:\/\/igormatias.com\/#website"},"datePublished":"2023-03-23T09:20:00+00:00","dateModified":"2026-02-02T13:04:43+00:00","description":"Abstract: Wearable devices are a useful and widely used source of continuous and temporal dependant data. In contrast...","breadcrumb":{"@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#breadcrumb"},"inLanguage":"pt-PT","potentialAction":[{"@type":"ReadAction","target":["https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/igormatias.com\/2023\/03\/23\/deja-vu-recurrent-neural-networks-for-health-wearables-data-forecast\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/igormatias.com\/"},{"@type":"ListItem","position":2,"name":"Dej\u00e0 vu: Recurrent Neural Networks for health wearables data forecast"}]},{"@type":"WebSite","@id":"https:\/\/igormatias.com\/#website","url":"https:\/\/igormatias.com\/","name":"Igor Matias","description":"(Computer Science, Health, AI)","publisher":{"@id":"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/igormatias.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"pt-PT"},{"@type":["Person","Organization"],"@id":"https:\/\/igormatias.com\/#\/schema\/person\/411aa904be362f1d407f8a491799a248","name":"Igor Matias","image":{"@type":"ImageObject","inLanguage":"pt-PT","@id":"https:\/\/igormatias.com\/#\/schema\/person\/image\/","url":"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg","contentUrl":"https:\/\/igormatias.com\/wp-content\/uploads\/2026\/04\/Matias_2026_square_933.jpg","width":933,"height":933,"caption":"Igor Matias"},"logo":{"@id":"https:\/\/igormatias.com\/#\/schema\/person\/image\/"},"description":"Ph.D. student in Information Systems at the University of Geneva, Switzerland, focusing Alzheimer\u2019s Disease diagnosis. Master of Computer Science and Engineering by University of Beira Interior, Portugal.","sameAs":["https:\/\/igormatias.com\/","https:\/\/www.facebook.com\/igormatias16","https:\/\/instagram.com\/igormatias3","https:\/\/linkedin.com\/in\/igormatias3","https:\/\/x.com\/IgorMatias3"]}]}},"_links":{"self":[{"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/posts\/774"}],"collection":[{"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/comments?post=774"}],"version-history":[{"count":4,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/posts\/774\/revisions"}],"predecessor-version":[{"id":1067,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/posts\/774\/revisions\/1067"}],"wp:attachment":[{"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/media?parent=774"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/categories?post=774"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/igormatias.com\/pt\/wp-json\/wp\/v2\/tags?post=774"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}