All Available Models

There are 72 available methods in the repository. For 57 methods, the equations/cut-points/formulas/etc. can be implemented in any software. For 15 methods, the models or functions are implemented in R.

• Ahmadi, 2020- four machine learning models to predict energy expenditure of preschoolers from raw acceleration features from hip or non-dominant wrist-worn ActiGraph accelerometers

• Aittasalo, 2015- cut-points for mean amplitude deviation in adolescents wearing Hookie or ActiGraph monitors at the hip

• Bammann, 2021- cut-points for Euclidean norm minus one for monitors worn at each hip, wrist, and ankle in older adults

• Bai, 2016-calculates the acceleration-based metric Activity Index from a hip-worn ActiGraph accelerometer which was related to energy expenditure in adult women

• Bianchim, 2022- cut-points for Euclidean norm minus one and mean amplitude deviation for monitors worn at the right hip and each wrist in children and adolescents with and without cystic fibrosisBrage, 2003- branched and not branched models to predict energy expenditure from hip-worn CSA accelerometer counts/min and/or heart rate in adult males

• Brandes, 2012- activity-specific linear regression equations for predicting energy expenditure during walking, cycling, and stair walking in children and adults using acceleration measured by a Dynaport on the lower back

• Brønd, 2019- six sets of cut-points for non-proprietary counts from children wearing an ActiGraph on the right hip

• Choi, 2010- five prediction models for energy expenditure in adolescents using ActiGraph counts from the hip, wrist, or ankle

• Chuang, 2013- single regression, activity-specific regression, and a mono-exponential equation to predict energy expenditure in adults wearing accelerometers on the wrist and ankle

• Clevenger, 2022- consensus method for time spent in moderate-to-vigorous physical activity using hip-worn ActiGraph data in adults

• Crotti, 2020- cut-points for Euclidean norm minus one from ActiGraph worn on right hip and each wrist in children

• Crouter, 2006- a 2-regression model for hip-worn ActiGraph counts in adults

• Crouter, 2007- a 2-regression model for hip-worn Actical counts in adults

• Crouter, 2010- an updated 2-regression model for hip-worn ActiGraph counts in adults

• Crouter, 2011- an updated 2-regression model for hip-worn Actical counts in adults

• Crouter, 2012- a 2-regression model for hip-worn ActiGraph counts in children

• Crouter, 2018- a 2-regression model for ankle-worn ActiGraph counts in children

• Curone, 2010- cut-points for signal magnitude area using ADXL330 on upper part of trunk

• Dibben, 2020- cut-points for two acceleration-based metrics (sum of vector magnitudes and mean amplitude deviation) for GENEActiv monitors worn at the hip and each wrist in patients with heart failure

• Dillon, 2016- cut-points for signal vector magnitude from GENEActiv worn on each wrist in adults

• Diniz-Sousa, 2020- regression equations and cut-points for Euclidean norm minus one, mean amplitude deviation, and vector magnitude of ActiGraph counts from ActiGraphs worn on the lower back or hip in adults with class 2 or 3 obesity

• Duclos, 2015- equation to predict total energy expenditure from smartphone accelerometers in adults

• Ellingson, 2016- modified Sojourn to predict energy expenditure of adults from hip-worn ActiGraph counts and thigh-worn ActivPal activity classification data

• Ellingson, 2017- modification of the Hildebrand et al. (2014) model to predict energy expenditure in adults using Euclidean norm minus one from an ActiGraph worn on the right wrist

• Esliger, 2011- cut-points for signal magnitude vector from GENEA worn on right hip and each wrist in adults

• Fraysse, 2021- cut-points for signal vector magnitude from GENEActiv worn on each wrist in older adults

• Heil, 2006- single and 2-regression models for predicting activity energy expenditure in children and adults wearing an Actical accelerometer on the right hip or non-dominant ankle or wrist

• Hibbing, 2018- four Sojourn models to predict energy expenditure of children using hip- or wrist-worn ActiGraph counts or Euclidean norm minus one

• Hibbing, 2018 (2)- fifteen 2-regression models to predict energy expenditure in adults for an ActiGraph worn at the hip, each wrist and ankle using Euclidean norm minus one, gyroscope vector magnitude, and direction changes

• Hikihara, 2014- a 2-regression model to predict energy expenditure in children wearing a waist-worn Omron accelerometer

• Hildebrand, 2014- calculates Euclidean norm minus one from raw acceleration data of hip or non-dominant wrist-worn ActiGraph or GENEActivs with energy expenditure equations and activity intensity cut-points for children and adults

• Hiremath, 2012- a general equation and a set of four activity-specific regression equations for predicting energy expenditure of manual wheelchair users using a Sensewear armband

• Horner, 2012- models for predicting total energy expenditure and physical activity energy expenditure for males and females during free-living using a 3dnx on the lower back

• Jang, 2006- combination of 7 accelerometers (15 total axes) is used to predict energy expenditure in adults using a simple calculation converting acceleration into velocity and work

• Jimmy, 2013- overall linear regression model, and a cubic 2-regression model and linear 2-regression model with separate equations for locomotor and play activities to predict energy expenditure in children wearing an ActiGraph on the hip

• Johansson, 2006- branched equation model to predict energy expenditure from heart rate and counts in adults wearing a ActiGraph on the lower back

• Kim, 2008- three linear regression models to predict energy expenditure in adults using waist, ankle, and wrist accelerometers

• Kiuchi, 2014- twelve total models (three feature sets at four wear locations including both wrists and upper arms) were developed for predicting energy expenditure using acceleration and gyroscopic angular velocity in manual wheelchair users

• Lyden, 2014- Sojourn model to predict energy expenditure from hip-worn ActiGraph count data in adults

• Mackintosh, 2016- thirteen machine learning models to predict energy expenditure in children using count data from ActiGraphs on chest, both hips, wrists, ankles, and knees

• Mehta, 2013- cut-points for activity intensity from raw acceleration

• Migueles, 2021- cut-points for activity intensity using Euclidean norm minus measured at the hip or either wrist in older adults

• Montoye, 2015- sixteen acceleration-based machine learning models to predict energy expenditure of adults wearing ActiGraph on thigh or hip or a GENEActiv on either wrist

• Montoye, 2016 (2)- four acceleration-based machine learning models to predict energy expenditure of adults wearing ActiGraph on thigh or hip or GENEActiv on either wrist

• Montoye, 2016 (3)- six acceleration-based machine learning models to predict energy expenditure of adults wearing GENEActiv on either wrist

• Montoye, 2017- four acceleration-based machine learning models to predict energy expenditure of adults wearing ActiGraph on ankle, hip, or either wrist

• Montoye, 2017 (2)- an artificial neural network to predict energy expenditure of adults wearing ActivPal on right thigh

• Montoye, 2018- four machine learning models to predict energy expenditure of adults wearing GENEActiv on left wrist

• Montoye, 2019- six machine learning models using either count or raw data from ActiGraphs on right hip or left wrist to predict energy expenditure of children

• Nguyen, 2013- data from three monitors - a Lifecorder accelerometer at the waist, Step Watch 3 at the ankle, and a GPS unit - are combined to predict energy expenditure in adults

• Nolan, 2014- acceleration measured by an iPod touch worn on the lower back is used to determine speed and grade of walking or running which are used to calculate energy expenditure based on pre-existing equations in adults

• O'Brien, 2021- logarithmic equation to predict METs from thigh-worn activPAL data in adults

• Ohkawara, 2011- magnitude of filtered acceleration and the ratio of unfiltered to filtered acceleration is used to determine activity type (sedentary, household, or locomotive), then activity-specific equations are used to predict energy expenditure in adults wearing a waist-worn Omron accelerometer

• Phillips, 2013- cut-points for signal magnitude vector from GENEA worn on right hip and each wrist in children

• Roscoe, 2017- cut-points for signal magnitude vector from GENEActiv worn on non-dominant wrist in preschool-aged children

• Sanders, 2019- cut-points for Euclidean norm minus one from ActiGraph worn on right hip and GENEActiv worn on non-dominant wrist in older adults

• Schaefer, 2014- cut-points for signal magnitude vector divided by sampling frequency from GENEActiv worn on non-dominant wrist in children

• Staudenmayer, 2015- linear regression and decision tree to predict energy expenditure or activity intensity, respectively, from dominant wrist-worn ActiGraphs in adults

• Steenbock, 2019- twenty-four machine learning models to predict energy expenditure of preschoolers using acceleration metrics from an ActiGraph on either hip, GENEActiv right hip, GENEActiv on either wrist, and/or ActivPal right thigh

• Tanaka, 2019- the ratio of unfiltered to filtered acceleration is used to determine non-locomotive vs. locomotive activities, then activity-specific equations are used to predict energy expenditure in young children wearing a waist-worn Omron accelerometer

• Tanaka, 2007- linear and non-linear equations were developed to predict energy expenditure in young children wearing an ActivTracer on the left hip

• Trost, 2016- two decision tree models for classifying activity intensity from ActiGraph counts in children with cerebral palsy wearing an ActiGraph at the hip

• Vaha-Ypya, 2015- equations to predict energy expenditure and activity intensity cut-points for mean amplitude deviation were determined from adults wearing a hip-worn Hookie monitor

• Vaha-Ypya, 2015 (2)- cut-points for mean amplitude deviation (MAD) that can be used for any accelerometer brand were determined from adults wearing three monitors at the hip

• van Hees, 2011- prediction equations for energy expenditure for pregnant women and non-pregnant women wearing a GENEA on the wrist or non-dominant hip

• van Hees, 2013- prediction equations for energy expenditure of women wearing a wrist-worn GENEA

• Wang, 2022- linear regression and cut-points for accelerometer activity index from a hip-worn ActiGraph monitor in older women

• Weippert, 2013- four models to predict energy expenditure in adults wearing a chest-worn accelerometer

• Yamazaki, 2009- model to predict energy expenditure during walking in adults wearing a waist-worn accelerometer

• Zakeri, 2010- a multivariate adaptive regression splines (MARS) model to predict 24-hour total energy expenditure overall and awake, sleep, and activity energy expenditure separately in children and adolescents wearing an Actiheart

• Zakeri, 2012- Cross-sectional time series and multivariate adaptive regression splines for ActiHeart, ActiGraph, or ActiGraph plus heart rate inputs, resulting in six total models for predicting total energy expenditure in preschool-aged children