Crossmark

Development and evaluation of wrist- and thigh-worn accelerometer algorithms using self-training machine learning models for classification of activity type and posture: towards device placement-agnostic methods in the ProPASS consortium

Crossref DOI link: https://doi.org/10.1186/s12966-026-01930-z

Published Online: 2026-07-02

Update policy: https://doi.org/10.1007/springer_crossmark_policy

Authors

Ahmadi, Matthew N.

Koemel, Nicholas

Biswas, Raaj

Holtermann, Andreas

Koster, Annemarie

Atkins, Andrew

Pulsford, Richard

del Pozo Cruz, Borja

Rangul, Vegar

Mitchell, John

Blodgett, Joanna

Hettiarachchi, Pasan

Svartengren, Magnus

Granat, Malcolm

Clark, Bronwyn

Hamer, Mark

Stamatakis, Emmanuel
Funding

Funding for this research was provided by:

National Heart Foundation of Australia (APP 107158)
License Information

Text and Data Mining valid from 2026-07-02

Version of Record valid from 2026-07-02
More Information

Article History

Received: 26 August 2025

Accepted: 27 April 2026

First Online: 2 July 2026

Declarations

:

: This study was approved by the Sydney Local Health District (ETH02815, ETH00138). Before participation, written consent was obtained from each participant.

: Not applicable.

: ES is a paid consultant and holds equity in Complement 1, a US-based company whose products and services relate to promotion of physical activity and other lifestyle behaviours. All other authors disclose no conflict of interest for this work.

Document is current

Any future updates will be listed below