Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Video

Updated: Apr 20, 2026

A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers
07:24

A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers

Published on: April 21, 2017

13.2K

Using hidden markov models to improve quantifying physical activity in accelerometer data - a simulation study.

Vitali Witowski1, Ronja Foraita2, Yannis Pitsiladis3

  • 1Department Biometry and Data Management, Leibniz Institute for Prevention Research and Epidemiology - BIPS, Bremen, Germany; Department of Mathematics and Computer Science, University of Bremen, Bremen, Germany.

Plos One
|December 3, 2014
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Physiological Demands and Contributors to Variance of Core Temperature in 162 Recreational Runners during Distance Running in a Warm-Humid Environment.

Medicine and science in sports and exercise·2026
Same author

Ascertainment of Cancer Cases in the German National Cohort (NAKO): Methods and Initial Results.

Deutsches Arzteblatt international·2026
Same author

Statistical methods for retrospective harmonization of longitudinal epidemiological data: a scoping review.

European journal of epidemiology·2026
Same author

Improving finite sample performance of causal discovery by exploiting temporal structure.

Statistical methods in medical research·2026
Same author

Real-time thermoregulatory and cardiovascular monitoring of non-acclimatised mountaineers in extreme cold: a 10-day field expedition study.

Frontiers in physiology·2026
Same author

ConNIS and labeling instability: New statistical methods for improving the detection of essential genes in TraDIS libraries.

PLoS computational biology·2026
This summary is machine-generated.

Hidden Markov Models (HMM) significantly improve physical activity (PA) classification from accelerometer data compared to traditional cutpoint methods. HMM with Gaussian distribution offers the most accurate results for objective PA measurement.

Area of Science:

  • Biomedical Engineering
  • Wearable Technology
  • Physical Activity Measurement

Background:

  • Accelerometers are the preferred method for objective physical activity (PA) measurement.
  • Traditional cutpoint methods classify PA based on device counts, but lack accuracy.
  • Hidden Markov Models (HMM) offer a novel approach to enhance PA classification.

Purpose of the Study:

  • To compare the accuracy of HMMs against the traditional cutpoint method for classifying PA from accelerometer data.
  • To evaluate different HMM distributions (Poisson, Generalized Poisson, Gaussian) for PA classification.
  • To assess HMM performance in detecting activity bouts and the number of activities.

Main Methods:

  • Simulated 1,000 days of labeled accelerometer data.

More Related Videos

Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity
05:59

Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity

Published on: March 7, 2019

7.4K
Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption
08:45

Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption

Published on: June 20, 2025

779

Related Experiment Videos

Last Updated: Apr 20, 2026

A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers
07:24

A Method for Quantifying Upper Limb Performance in Daily Life Using Accelerometers

Published on: April 21, 2017

13.2K
Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity
05:59

Visualization of Intensity Levels to Reduce the Gap Between Self-Reported and Directly Measured Physical Activity

Published on: March 7, 2019

7.4K
Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption
08:45

Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption

Published on: June 20, 2025

779
  • Compared traditional cutpoint method with HMMs (Poisson, Generalized Poisson, Gaussian).
  • Evaluated misclassification rate (MCR), bout detection, activity count detection, and runtime.

    • Main Results:

    • HMMs demonstrated lower MCR than cutpoints: HMM[Gauss] <2%, HMM[GenPois] 3%, HMM[Pois] 8%, cutpoint 11%.
    • HMMs showed improved detection of activity bouts and number of activities compared to cutpoints.
    • HMM[Gauss] achieved the lowest MCR, while HMM[GenPois] better identified the number of activities.

    Conclusions:

    • HMM-based methods significantly outperform the traditional cutpoint method for accelerometer-based PA classification.
    • HMMs are suitable for modeling accelerometer data and improving objective PA assessment.
    • HMM with Gaussian distribution is recommended as the most appropriate method for real-life accelerometer data analysis.