Search research articles

Search research articles

Related Experiment Video

Updated: Jun 16, 2026

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Published on: December 15, 2023

Applying Multivariate Segmentation Methods to Human Activity Recognition From Wearable Sensors' Data.

Kenan Li¹, Rima Habre¹, Huiyu Deng¹

¹Department of Preventive Medicine, Keck School of Medicine of University of Southern California, Los Angeles, CA, United States.

JMIR Mhealth and Uhealth

|February 8, 2019

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

Electric vehicle adoption, gentrification, and housing prices: a longitudinal study in California.

Environmental research communications·2026

Same author

Developing Scenario-Based Strategies for Health, Climate, and Environmental Preparedness: The One Health, One Earth Approach.

GeoHealth·2026

Same author

TGF-β Pathway-Based Polygenic Risk Score Modifies the Association between Red Meat Intake and Colorectal Cancer Risk: Application of a Novel Pathway-Based PRS Method.

Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology·2026

Same author

Translational framework linking perfluoroheptanoic acid (PFHpA) exposure to metabolic dysfunction associated steatotic liver disease in adolescents.

Communications medicine·2025

Same author

Plasma Metabolomic and Proteomic Signatures of Blood Pressure Management After Bariatric Surgery Among Adolescents.

Hypertension (Dallas, Tex. : 1979)·2025

Same author

Dust Events and Children's Lung Function Near a Drying Saline Lake.

American journal of respiratory and critical care medicine·2025

Same journal

Alleviating Nurse Burnout With an Artificial Intelligence-Selected Mobile Cognitive Behavioral Therapy-Based Intervention: Mixed Methods Randomized Controlled Trial.

JMIR mHealth and uHealth·2026

Same journal

Efficacy of a Cognitive Behavioral Therapy-Based Online Self-Help Group for Depression and Suicide Ideation: Randomized Controlled Trial.

JMIR mHealth and uHealth·2026

Same journal

Inclusive Contactless Monitoring for Older Adults From Diverse Backgrounds: Mixed Methods Study.

JMIR mHealth and uHealth·2026

Same journal

Systematic and Collaborative Approach to Learning and Educational Content Development (SCALED) for Health Apps: An Experience-Informed Conceptual Framework.

JMIR mHealth and uHealth·2026

Same journal

Effects of Telerehabilitation Based on Motion Recognition Technology on Exercise Endurance of Patients With Non-Small Cell Lung Cancer After Surgery: Single-Center, Prospective, Open-Label, Randomized Controlled Trial.

JMIR mHealth and uHealth·2026

Same journal

Feasibility and Acceptability of mPallCare<i>,</i> a Digital Health Intervention for People Living With Advanced Cancer in a Refugee Settlement in Uganda: Mixed Methods Study.

JMIR mHealth and uHealth·2026

See all related articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

This summary is machine-generated.

Greedy Gaussian segmentation (GGS) offers improved human activity recognition (HAR) by adapting to variable activity durations. This method provides more stable predictions and higher accuracy than fixed-size windows, aiding personalized medicine and research.

Area of Science:

Biomedical engineering
Machine learning
Wearable technology

Background:

Accurate physical activity quantification is crucial for personalized medicine and epidemiological studies, such as managing asthma exacerbations.
Existing human activity recognition (HAR) algorithms often use fixed-size windows, which struggle with real-world variable activity durations.
Short windows yield noisy data, while long windows miss brief, intense activities.

Purpose of the Study:

To develop an HAR framework that accommodates variable activity bout durations.
To detect change points in multivariate time series data to define activity bouts.
To predict physical activity within these dynamically defined, homogeneous windows.

Main Methods:

Applied fixed-width sliding windows and greedy Gaussian segmentation (GGS) to sensor data (accelerometer, gyroscope).

Keywords:

machine learning physical activity smartphone statistical data analysis wearable devices

More Related Videos

Design and Analysis for Fall Detection System Simplification

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

Related Experiment Videos

Last Updated: Jun 16, 2026

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Artificial Intelligence-Based System for Detecting Attention Levels in Students

Published on: December 15, 2023

Design and Analysis for Fall Detection System Simplification

Design and Analysis for Fall Detection System Simplification

Published on: April 6, 2020

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment

Published on: December 11, 2015

Utilized Xgboost models for activity prediction within segmented windows.

Validated methods on the HARuS (adults, smartphone) and BREATHE (children, smartwatch) datasets, including simulated variable activity durations.

Main Results:

GGS produced the least noisy predictions in both datasets.
GGS achieved high accuracy rates: 91.06% in HARuS and 79.4% in BREATHE.
While fixed-size windows showed slightly higher peak accuracy in HARuS, GGS demonstrated superior performance with variable bout durations.

Conclusions:

GGS multivariate segmentation effectively creates adaptive windows for more stable and accurate HAR compared to fixed-size methods.
The approach shows promise for real-world applications, though accuracy varies with data collection context (e.g., device, population).
GGS, initially offline, is adaptable for real-time HAR prediction using streaming data.