A Machine Learning Approach for Physical Activity Recognition in Cystic Fibrosis - Research report - results on accelerometer-based prediction

Aug 5, 2026

Upload by Finn

A Machine Learning Approach for Physical Activity Recognition in Cystic Fibrosis develops and validates machine learning models to predict physical activity (PA) intensities in children and adolescents with cystic fibrosis (CF) across different accelerometry brands and placement locations. The study includes 35 participants with CF and 28 healthy youth who perform six activities while wearing GENEActiv devices (both wrists) and ActiGraph GT9X devices (both wrists and waist). Three supervised classifiers—K-Nearest Neighbour, Random Forest, and eXtreme Gradient Boosted Decision Trees—use 10-fold cross-validation to identify input signal patterns. Results show that most models achieve over 97% accuracy with sensitivity and specificity above 95% regardless of brand, placement, or health condition, while dominant-wrist/waist configurations fail for vigorous intensity PA prediction.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Abstract

Introduction

Background on CF and activity limitations

Limitations of traditional accelerometer cut-points

Rationale for machine learning methods

Measurement in Physical Education and Exercise Science ISSN: (Print) (Online) Journal homepage: [https://www.tandfonline.com/loi/hmpe20](https://www.tandfonline.com/loi/hmpe20) A Machine Learning Approach for Physical Activity Recognition in Cystic Fibrosis Mayara S. Bianchim, Melitta A. McNarry, Alan R. Barker, Craig A. Williams, Sarah Denford, Lena Thia, Rachel Evans & Kelly A Mackintoshon behalf of ActiveYouth SRC group To cite this article: Mayara S. Bianchim, Melitta A. McNarry, Alan R. Barker, Craig A. Williams, Sarah Denford, Lena Thia, Rachel Evans & Kelly A Mackintoshon behalf of ActiveYouth SRC group (24 Oct 2023): A Machine Learning Approach for Physical Activity Recognition in Cystic Fibrosis, Measurement in Physical Education and Exercise Science, DOI: 10. 1080/1091367X.2023.2271444 To link to this article: [https://doi.org/10.1080/1091367X.2023.2271444](https://doi.org/10.1080/1091367X.2023.2271444) © 2023 The Author(s) . Published with license by Taylor & Francis Group, LLC. | View supplementary material | | --- | | Published online: 24 Oct 2023. | | Submit your article to this journal | | Article views: 98 | | View related articles | | View Crossmark data | Full Terms & Conditions of access and use can be found at [https://www.tandfonline.com/action/journalInformation?journalCode=hmpe20](https://www.tandfonline.com/action/journalInformation?journalCode=hmpe20) MEASUREMENT IN PHYSICAL EDUCATION AND EXERCISE SCIENCE [https://doi.org/10.1080/1091367X.2023.2271444](https://doi.org/10.1080/1091367X.2023.2271444) A Machine Learning Approach for Physical Activity Recognition in Cystic Fibrosis Mayara S. Bianchima,b, Melitta A. McNarry a, Alan R. Barkerc, Craig A. Williamsc, Sarah Denfordc, Lena Thiad, Rachel Evanse, Kelly A Mackintosh a, and on behalf of ActiveYouth SRC group aApplied Sports, Technology, Exercise and Medicine Research Centre, Swansea University, Swansea, UK; bSchool of Medical and Health Sciences, Bangor University, Bangor, UK; cChildren’s Health and Exercise Research Centre, University of Exeter, Exeter, UK; dDepartment of Paediatric Respiratory Medicine and Cystic Fibrosis Unit, Noah’s Ark Children’s Hospital for Wales, Cardiff, UK; ePaediatric Department, Morriston Hospital, Swansea, UK ABSTRACT This study aimed to develop and validate machine learning models to predict intensities in children and adolescents with cystic fibrosis (CF) across different accelerometry brands and placements. Thirty-five children and adolescents with CF (11.6 ± 2.8 yrs; 15 girls) and 28 healthy youth (12.2 ± 2.7 yrs; 16 girls) performed six activities whilst wearing GENEActivs (both wrists) and ActiGraphs GT9X (both wrists and waist) . Three supervised learning classifiers (K-Nearest Neighbour, Random Forest and eXtreme Gradient Boosted Decision Tree) were used to identify the input signal pattern for each PA type and intensity, with a 10-fold cross-validation utilized to assess the performance of the classifiers. ActiGraph GT9X on the dominant wrist and waist and GENEActiv on the dominant wrist failed to predict vigorous intensity PA activities. All other models, for activity type and intensities, exceeded 97% accuracy, with a sensitivity and specificity of greater than 95%, irrespective of accelerometer brand, placement or health condition. KEYWORDS Threshold; Physical Activity; ENMO; MAD; youth Introduction Cystic Fibrosis (CF) is the most common life-limiting autosomal recessive disorder in the Caucasian population affecting over 10,500 individuals in the United Kingdom (Trust, 2018) . Limited exercise tolerance in CF is multifactorial and can promote physical inactivity which has significant negative health implications (González et al., 2017) . Habitual physical activity (PA), and particularly moderate-to-vigorous physical activity (MVPA), is associated with significant health benefits in youth with CF (Hebestreit et al., 2014), including an increased lifespan and b

FAQs

What was the main goal of the study on physical activity recognition in cystic fibrosis?

To develop and validate machine learning models that predict PA intensities in children and adolescents with cystic fibrosis across different accelerometer brands and placements.

Which accelerometers and placements were used during data collection?

Participants wore GENEActiv on both wrists and ActiGraph GT9X on both wrists and the waist while performing six activities.

How were the machine learning classifiers evaluated?

Three supervised classifiers were assessed using 10-fold cross-validation to determine their ability to identify PA types and intensity levels.

Aug 5, 2026

Upload by Finn