scholarly journals Towards a Portable Model to Discriminate Activity Clusters from Accelerometer Data

Sensors ◽  
2019 ◽  
Vol 19 (20) ◽  
pp. 4504 ◽  
Author(s):  
Petra Jones ◽  
Evgeny M. Mirkes ◽  
Tom Yates ◽  
Charlotte L. Edwardson ◽  
Mike Catt ◽  
...  
Keyword(s):  
Physical Activity ◽  
Cross Validation ◽  
Accelerometer Data ◽  
Activity Type ◽  
Free Living ◽  
Independent Dataset ◽  
Clustering Model ◽  
Activity Types ◽  
Unsupervised Approach ◽  
Viable Approach

Few methods for classifying physical activity from accelerometer data have been tested using an independent dataset for cross-validation, and even fewer using multiple independent datasets. The aim of this study was to evaluate whether unsupervised machine learning was a viable approach for the development of a reusable clustering model that was generalisable to independent datasets. We used two labelled adult laboratory datasets to generate a k-means clustering model. To assess its generalised application, we applied the stored clustering model to three independent labelled datasets: two laboratory and one free-living. Based on the development labelled data, the ten clusters were collapsed into four activity categories: sedentary, standing/mixed/slow ambulatory, brisk ambulatory, and running. The percentages of each activity type contained in these categories were 89%, 83%, 78%, and 96%, respectively. In the laboratory independent datasets, the consistency of activity types within the clusters dropped, but remained above 70% for the sedentary clusters, and 85% for the running and ambulatory clusters. Acceleration features were similar within each cluster across samples. The clusters created reflected activity types known to be associated with health and were reasonably robust when applied to diverse independent datasets. This suggests that an unsupervised approach is potentially useful for analysing free-living accelerometer data.

scholarly journals A Machine Learning Classifier for Detection of Physical Activity Types and Postures During Free-Living

2021 ◽  
pp. 1-8
Author(s):  
Kerstin Bach ◽  
Atle Kongsvold ◽  
Hilde Bårdstu ◽  
Ellen Marie Bardal ◽  
Håkon S. Kjærnli ◽  
...  
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Ground Truth ◽  
Daily Physical Activity ◽  
Gradient Boosting ◽  
Free Living ◽  
Recording Time ◽  
Learning Classifier ◽  
Extreme Gradient Boosting ◽  
Activity Types

Introduction: Accelerometer-based measurements of physical activity types are commonly used to replace self-reports. To advance the field, it is desirable that such measurements allow accurate detection of key daily physical activity types. This study aimed to evaluate the performance of a machine learning classifier for detecting sitting, standing, lying, walking, running, and cycling based on a dual versus single accelerometer setups during free-living. Methods: Twenty-two adults (mean age [SD, range] 38.7 [14.4, 25–68] years) were wearing two Axivity AX3 accelerometers positioned on the low back and thigh along with a GoPro camera positioned on the chest to record lower body movements during free-living. The labeled videos were used as ground truth for training an eXtreme Gradient Boosting classifier using window lengths of 1, 3, and 5 s. Performance of the classifier was evaluated using leave-one-out cross-validation. Results: Total recording time was ∼38 hr. Based on 5-s windowing, the overall accuracy was 96% for the dual accelerometer setup and 93% and 84% for the single thigh and back accelerometer setups, respectively. The decreased accuracy for the single accelerometer setup was due to a poor precision in detecting lying based on the thigh accelerometer recording (77%) and standing based on the back accelerometer recording (64%). Conclusion: Key daily physical activity types can be accurately detected during free-living based on dual accelerometer recording, using an eXtreme Gradient Boosting classifier. The overall accuracy decreases marginally when predictions are based on single thigh accelerometer recording, but detection of lying is poor.

scholarly journals Reexamination of Accelerometer Calibration with Energy Expenditure as Criterion: VO2net Instead of MET for Age-Equivalent Physical Activity Intensity

Sensors ◽  
2019 ◽  
Vol 19 (15) ◽  
pp. 3377 ◽  
Author(s):  
Daniel Arvidsson ◽  
Jonatan Fridolfsson ◽  
Christoph Buck ◽  
Örjan Ekblom ◽  
Elin Ekblom-Bak ◽  
...  
Keyword(s):  
Physical Activity ◽  
Age Groups ◽  
Accelerometer Data ◽  
Physical Activity Intensity ◽  
Free Living ◽  
Activity Intensity ◽  
Age Related ◽  
Younger Age ◽  
Metabolic Equivalent Of Task ◽  
Adolescents And Adults

Accelerometer calibration for physical activity (PA) intensity is commonly performed using Metabolic Equivalent of Task (MET) as criterion. However, MET is not an age-equivalent measure of PA intensity, which limits the use of MET-calibrated accelerometers for age-related PA investigations. We investigated calibration using VO2net (VO2gross − VO2stand; mL⋅min−1⋅kg−1) as criterion compared to MET (VO2gross/VO2rest) and the effect on assessment of free-living PA in children, adolescents and adults. Oxygen consumption and hip/thigh accelerometer data were collected during rest, stand and treadmill walk and run. Equivalent speed (Speedeq) was used as indicator of the absolute speed (Speedabs) performed with the same effort in individuals of different body size/age. The results showed that VO2net was higher in younger age-groups for Speedabs, but was similar in the three age-groups for Speedeq. MET was lower in younger age-groups for both Speedabs and Speedeq. The same VO2net-values respective MET-values were applied to all age-groups to develop accelerometer PA intensity cut-points. Free-living moderate-and-vigorous PA was 216, 115, 74 and 71 min/d in children, adolescents, younger and older adults with VO2net-calibration, but 140, 83, 74 and 41 min/d with MET-calibration, respectively. In conclusion, VO2net calibration of accelerometers may provide age-equivalent measures of PA intensity/effort for more accurate age-related investigations of PA in epidemiological research.

scholarly journals Reallocation of time between device-measured movement behaviours and risk of incident cardiovascular disease

2021 ◽  
pp. bjsports-2021-104050
Author(s):  
Rosemary Walmsley ◽  
Shing Chan ◽  
Karl Smith-Byrne ◽  
Rema Ramakrishnan ◽  
Mark Woodward ◽  
...  
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Sedentary Behaviour ◽  
Lower Risk ◽  
Accelerometer Data ◽  
Uk Biobank ◽  
Free Living ◽  
Machine Learning Methods ◽  
Machine Learning Model ◽  
Incident Cardiovascular Disease

ObjectiveTo improve classification of movement behaviours in free-living accelerometer data using machine-learning methods, and to investigate the association between machine-learned movement behaviours and risk of incident cardiovascular disease (CVD) in adults.MethodsUsing free-living data from 152 participants, we developed a machine-learning model to classify movement behaviours (moderate-to-vigorous physical activity behaviours (MVPA), light physical activity behaviours, sedentary behaviour, sleep) in wrist-worn accelerometer data. Participants in UK Biobank, a prospective cohort, were asked to wear an accelerometer for 7 days, and we applied our machine-learning model to classify their movement behaviours. Using compositional data analysis Cox regression, we investigated how reallocating time between movement behaviours was associated with CVD incidence.ResultsIn leave-one-participant-out analysis, our machine-learning method classified free-living movement behaviours with mean accuracy 88% (95% CI 87% to 89%) and Cohen’s kappa 0.80 (95% CI 0.79 to 0.82). Among 87 498 UK Biobank participants, there were 4105 incident CVD events. Reallocating time from any behaviour to MVPA, or reallocating time from sedentary behaviour to any behaviour, was associated with lower CVD risk. For an average individual, reallocating 20 min/day to MVPA from all other behaviours proportionally was associated with 9% (95% CI 7% to 10%) lower risk, while reallocating 1 hour/day to sedentary behaviour from all other behaviours proportionally was associated with 5% (95% CI 3% to 7%) higher risk.ConclusionMachine-learning methods classified movement behaviours accurately in free-living accelerometer data. Reallocating time from other behaviours to MVPA, and from sedentary behaviour to other behaviours, was associated with lower risk of incident CVD, and should be promoted by interventions and guidelines.

scholarly journals An artificial neural network to estimate physical activity energy expenditure and identify physical activity type from an accelerometer

2009 ◽  
Vol 107 (4) ◽  
pp. 1300-1307 ◽  
Author(s):  
John Staudenmayer ◽  
David Pober ◽  
Scott Crouter ◽  
David Bassett ◽  
Patty Freedson
Keyword(s):  
Physical Activity ◽  
Oxygen Consumption ◽  
Confidence Interval ◽  
Mean Squared Error ◽  
Temporal Dynamics ◽  
Accelerometer Data ◽  
Ann Model ◽  
Activity Type ◽  
Metabolic Equivalents ◽  
Artificial Neural

The aim of this investigation was to develop and test two artificial neural networks (ANN) to apply to physical activity data collected with a commonly used uniaxial accelerometer. The first ANN model estimated physical activity metabolic equivalents (METs), and the second ANN identified activity type. Subjects ( n = 24 men and 24 women, mean age = 35 yr) completed a menu of activities that included sedentary, light, moderate, and vigorous intensities, and each activity was performed for 10 min. There were three different activity menus, and 20 participants completed each menu. Oxygen consumption (in ml·kg−1·min−1) was measured continuously, and the average of minutes 4–9 was used to represent the oxygen cost of each activity. To calculate METs, activity oxygen consumption was divided by 3.5 ml·kg−1·min−1 (1 MET). Accelerometer data were collected second by second using the Actigraph model 7164. For the analysis, we used the distribution of counts (10th, 25th, 50th, 75th, and 90th percentiles of a minute's second-by-second counts) and temporal dynamics of counts (lag, one autocorrelation) as the accelerometer feature inputs to the ANN. To examine model performance, we used the leave-one-out cross-validation technique. The ANN prediction of METs root-mean-squared error was 1.22 METs (confidence interval: 1.14–1.30). For the prediction of activity type, the ANN correctly classified activity type 88.8% of the time (confidence interval: 86.4–91.2%). Activity types were low-level activities, locomotion, vigorous sports, and household activities/other activities. This novel approach of applying ANNs for processing Actigraph accelerometer data is promising and shows that we can successfully estimate activity METs and identify activity type using ANN analytic procedures.

Validity of the Acti4 method for detection of physical activity types in free-living settings: comparison with video analysis

Ergonomics ◽  
2015 ◽  
Vol 58 (6) ◽  
pp. 953-965 ◽  
Author(s):  
Ingunn Stemland ◽  
Jørgen Ingebrigtsen ◽  
Caroline S. Christiansen ◽  
Bente R. Jensen ◽  
Christiana Hanisch ◽  
...  
Keyword(s):  
Physical Activity ◽  
Video Analysis ◽  
Free Living ◽  
Activity Types

scholarly journals Employing Machine Learning to Estimate Hallmark Measures of Physical Activities from Wrist-worn Devices Across Age Groups

2021 ◽  
Author(s):  
Mamoun T. Mardini ◽  
Chen Bai ◽  
Amal A. Wanigatunga ◽  
Santiago Saldana ◽  
Ramon Casanova ◽  
...  
Keyword(s):  
Physical Activity ◽  
Machine Learning ◽  
Energy Expenditure ◽  
Age Groups ◽  
Physical Activities ◽  
Accelerometer Data ◽  
Activity Type ◽  
Physical Activity Intensity ◽  
Activity Intensity ◽  
Score Range

Wrist-worn fitness trackers and smartwatches are proliferating with an incessant attention towards health tracking. Given the growing popularity of wrist-worn devices across all age groups, a rigorous evaluation for recognizing hallmark measures of physical activities and estimating energy expenditure is needed to compare their accuracy across the lifespan. The goal of the study was to build machine learning models to recognize physical activity type (sedentary, locomotion, and lifestyle) and intensity (low, light, and moderate), identify individual physical activities, and estimate energy expenditure. The primary aim of this study was to build and compare models for different age groups: young [20-50 years], middle (50-70 years], and old (70-89 years]. Participants (n = 253, 62% women, aged 20-89 years old) performed a battery of 33 daily activities in a standardized laboratory setting while wearing a portable metabolic unit to measure energy expenditure that was used to gauge metabolic intensity. Tri-axial accelerometer collected data at 80-100 Hz from the right wrist that was processed for 49 features. Results from random forests algorithm were quite accurate in recognizing physical activity type, the F1-Score range across age groups was: sedentary [0.955 – 0.973], locomotion [0.942 – 0.964], and lifestyle [0.913 – 0.949]. Recognizing physical activity intensity resulted in lower performance, the F1-Score range across age groups was: sedentary [0.919 – 0.947], light [0.813 – 0.828], and moderate [0.846 – 0.875]. The root mean square error range was [0.835 – 1.009] for the estimation of energy expenditure. The F1-Score range for recognizing individual physical activities was [0.263 – 0.784]. Performances were relatively similar and the accelerometer data features were ranked similarly between age groups. In conclusion, data features derived from wrist worn accelerometers lead to high-moderate accuracy estimating physical activity type, intensity and energy expenditure and are robust to potential age-differences.

scholarly journals Measuring Free-Living Physical Activity With Three Commercially Available Activity Monitors for Telemonitoring Purposes: Validation Study (Preprint)

2018 ◽  
Author(s):  
Martine JM Breteler ◽  
Joris H Janssen ◽  
Wilko Spiering ◽  
Cor J Kalkman ◽  
Wouter W van Solinge ◽  
...  
Keyword(s):  
Physical Activity ◽  
Validation Study ◽  
Cross Validation ◽  
Usability Evaluation ◽  
Step Count ◽  
Added Value ◽  
Reference Standard ◽  
Free Living ◽  
Activity Trackers ◽  
Leave One Out

BACKGROUND Remote monitoring of physical activity in patients with chronic conditions could be useful to offer care professionals real-time assessment of their patient’s daily activity pattern to adjust appropriate treatment. However, the validity of commercially available activity trackers that can be used for telemonitoring purposes is limited. OBJECTIVE The purpose of this study was to test usability and determine the validity of 3 consumer-level activity trackers as a measure of free-living activity. METHODS A usability evaluation (study 1) and validation study (study 2) were conducted. In study 1, 10 individuals wore one activity tracker for a period of 30 days and filled in a questionnaire on ease of use and wearability. In study 2, we validated three selected activity trackers (Apple Watch, Misfit Shine, and iHealth Edge) and a fourth pedometer (Yamax Digiwalker) against the reference standard (Actigraph GT3X) in 30 healthy participants for 72 hours. Outcome measures were 95% limits of agreement (LoA) and bias (Bland-Altman analysis). Furthermore, median absolute differences (MAD) were calculated. Correction for bias was estimated and validated using leave-one-out cross validation. RESULTS Usability evaluation of study 1 showed that iHealth Edge and Apple Watch were more comfortable to wear as compared with the Misfit Flash. Therefore, the Misfit Flash was replaced by Misfit Shine in study 2. During study 2, the total number of steps of the reference standard was 21,527 (interquartile range, IQR 17,475-24,809). Bias and LoA for number of steps from the Apple Watch and iHealth Edge were 968 (IQR −5478 to 7414) and 2021 (IQR −4994 to 9036) steps. For Misfit Shine and Yamax Digiwalker, bias was −1874 and 2004, both with wide LoA of (13,869 to 10,121) and (−10,932 to 14,940) steps, respectively. The Apple Watch noted the smallest MAD of 7.7% with the Actigraph, whereas the Yamax Digiwalker noted the highest MAD (20.3%). After leave-one-out cross validation, accuracy estimates of MAD of the iHealth Edge and Misfit Shine were within acceptable limits with 10.7% and 11.3%, respectively. CONCLUSIONS Overall, the Apple Watch and iHealth Edge were positively evaluated after wearing. Validity varied widely between devices, with the Apple Watch being the most accurate and Yamax Digiwalker the least accurate for step count in free-living conditions. The iHealth Edge underestimates number of steps but can be considered reliable for activity monitoring after correction for bias. Misfit Shine overestimated number of steps and cannot be considered suitable for step count because of the low agreement. Future studies should focus on the added value of remotely monitoring activity patterns over time in chronic patients.

Standardizing and Optimizing the Use of Accelerometer Data for Free-Living Physical Activity Monitoring

2005 ◽  
Vol 2 (3) ◽  
pp. 366-383 ◽  
Author(s):  
Dale W. Esliger ◽  
Jennifer L. Copeland ◽  
Joel D. Barnes ◽  
Mark S. Tremblay
Keyword(s):  
Public Health ◽  
Physical Activity ◽  
Data Capture ◽  
Accelerometer Data ◽  
Free Living ◽  
Physical Activity Monitoring ◽  
Technological Advances ◽  
Standardized Reporting ◽  
Public Health Officials ◽  
Reporting Procedures

The unequivocal link between physical activity and health has prompted researchers and public health officials to search for valid, reliable, and logistically feasible tools to measure and quantify free-living physical activity. Accelerometers hold promise in this regard. Recent technological advances have led to decreases in both the size and cost of accelerometers while increasing functionality (e.g., greater memory, waterproofing). A lack of common data reduction and standardized reporting procedures dramatically limit their potential, however. The purpose of this article is to expand on the utility of accelerometers for measuring free-living physical activity. A detailed example profile of physical activity is presented to highlight the potential richness of accelerometer data. Specific recommendations for optimizing and standardizing the use of accelerometer data are provided with support from specific examples. This descriptive article is intended to advance and ignite scholarly dialogue and debate regarding accelerometer data capture, reduction, analysis, and reporting.

Sign in / Sign up

Export Citation Format

Share Document