Methods to Estimate Aspects of Physical Activity and Sedentary Behavior from High-Frequency Wrist Accelerometer Measurements
Staudenmayer, 2015
Embedded Files
Last update: January 15, 2022. Updates will occur every 3 months.
Description
Description
Linear regression and decision tree to predict METs or activity intensity from acceleration data captured by dominant wrist-worn ActiGraphs in adults. The decision tree was 74% correct for classifying activity intensity while the linear regression had a root mean square error of approximately 1.55 METs.
Development/Validation
Development/Validation
Sample: 20 (10 males) adults, 20-39 years of age
Setting: Laboratory
Activities: Activities of daily living, treadmill walking and running, sports, stairs, stationary
Criterion: Oxycon Mobile (VO2)
Accelerometer(s): ActiGraph GT3X+ on dominant wrist
Validation approach: Leave-one-out cross-validation
Instructions
Instructions
Calculate sdvm as the standard deviation of the vector magnitude (square root of the sum of the squared acceleration in each axis) for the epoch/interval. Calculate mangle as the mean angle of acceleration relative to vertical on the device for the epoch/interval, where angle is equal to 90arcsin(x/vm)/(pi/2). Staudenmayer provides access to two of five described models: 1) a linear regression model to predict METs from sdvm and mangle and 2) a decision tree to classify activity intensity from sdvm and mangle.
Source Information
Source Information
Reference:
Staudenmayer, J., He, S., Hickey, A., Sasaki, J., & Freedson, P. (2015). Methods to estimate aspects of physical activity and sedentary behavior from high-frequency wrist accelerometer measurements. Journal of Applied Physiology, 119(4), 396-403. https://doi.org/10.1152/japplphysiol.00026.2015 Link to paper
Corresponding author: John Staudenmayer, jstauden@math.umass.edu
Contact
Contact
Kimberly Clevenger at accelerometerrepository@gmail.com
Page updated
Google Sites
Report abuse