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: Mar 1, 2026

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.5K

Modeling long-term human activeness using recurrent neural networks for biometric data.

Zae Myung Kim1, Hyungrai Oh2, Han-Gyu Kim1

  • 1School of Computing, KAIST, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, South Korea.

BMC Medical Informatics and Decision Making
|May 26, 2017
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

Deflationary Extraction Transformer for Speech Separation with Unknown Number of Talkers.

Sensors (Basel, Switzerland)·2025
Same author

Risk Factors Associated with Progression to Surgery in Patients with Ischemic Colitis.

The Korean journal of gastroenterology = Taehan Sohwagi Hakhoe chi·2024
Same author

Modulation of Phonons and Excitons Due to Moiré Potentials in Twisted Bilayer of WSe<sub>2</sub>/MoSe<sub>2</sub>.

ACS nano·2023
Same author

Multi-task learning approach for utilizing temporal relations in natural language understanding tasks.

Scientific reports·2023
Same author

Classification of Diffuse Glioma Subtype from Clinical-Grade Pathological Images Using Deep Transfer Learning.

Sensors (Basel, Switzerland)·2021
Same author

γ-GeSe: A New Hexagonal Polymorph from Group IV-VI Monochalcogenides.

Nano letters·2021

This study shows that user activeness, defined by biometric data, can be modeled and predicted using machine learning. Forecasting long-term activeness is feasible and can enhance health applications.

Area of Science:

  • * Biomedical engineering and data science.
  • * Machine learning applications in health and fitness monitoring.

Background:

  • * Fitness trackers enable continuous monitoring of user biometric data, including heart rate, footsteps, and calories burned.
  • * This data, termed 'activeness,' presents an opportunity for long-term user behavior analysis and prediction.

Purpose of the Study:

  • * To investigate the feasibility of modeling and predicting long-term user activeness.
  • * To evaluate the performance of various machine learning models in predicting activeness.
  • * To assess the capability of predicting periods of low user activeness.

Main Methods:

  • * Utilized several months of time-series biometric data from seven users.
  • * Proposed and compared four recurrent neural network (RNN) architectures, a deep neural network, and a regression model.
Keywords:
Activeness predictionCalorieFootstepHeart rateRecurrent neural networkTime series modeling

More Related Videos

Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

772

Related Experiment Videos

Last Updated: Mar 1, 2026

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention
06:37

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

5.5K
Decoding Natural Behavior from Neuroethological Embedding
08:00

Decoding Natural Behavior from Neuroethological Embedding

Published on: October 3, 2025

772
  • * Investigated hyper-parameter settings related to time-series length and predicted low activeness periods.

    • Main Results:

    • * Activeness data exhibited short-term autocorrelation; calories and footsteps were positively correlated, while heart rate showed minimal correlation.
    • * Recurrent neural network (RNN) models performed best, though differences with baseline models like linear regression were marginal.
    • * Accurate prediction of future low activeness periods (e.g., 84% precision for the next hour) was achieved using trained RNN models.

    Conclusions:

    • * The concept of user 'activeness' was defined and its long-term forecasting was demonstrated as feasible.
    • * Predictive models can forecast future user activeness levels.
    • * This capability can be integrated into health applications for proactive user recommendations.