A random forest classifier for the prediction of energy expenditure and type of physical activity from wrist and hip accelerometers

Abstract

Wrist accelerometers are being used in population level surveillance of physical activity (PA) but more research is needed to evaluate their validity for correctly classifying types of PA behavior and predicting energy expenditure (EE). In this study we compare accelerometers worn on the wrist and hip, and the added value of heart rate (HR) data, for predicting PA type and EE using machine learning. Forty adults performed locomotion and household activities in a lab setting while wearing three ActiGraph GT3X+ accelerometers (left hip, right hip, non-dominant wrist) and a HR monitor (Polar RS400). Participants also wore a portable indirect calorimeter (COSMED K4b2), from which EE and metabolic equivalents (METs) were computed for each minute. We developed two predictive models: a random forest classifier to predict activity type and a random forest of regression trees to estimate METs. Predictions were evaluated using leave-one-user-out cross-validation. The hip accelerometer obtained an average accuracy of 92.3% in predicting four activity types (household, stairs, walking, running), while the wrist accelerometer obtained an average accuracy of 87.5%. Across all 8 activities combined (laundry, window washing, dusting, dishes, sweeping, stairs, walking, running), the hip and wrist accelerometers obtained average accuracies of 70.2% and 80.2% respectively. Predicting METs using the hip or wrist devices alone obtained root mean square errors (rMSE) of 1.09 and 1.00 METs per 6-minute bout, respectively. Including HR data improved MET estimation, but did not significantly improve activity type classification. These results demonstrate the validity of random forest classification and regression forests for PA type and MET prediction using accelerometers. The wrist accelerometer proved more useful in predicting activities with significant arm movement, while the hip accelerometer was superior for predicting locomotion and estimating EE.

Intracranial Pressure Forecasting in Children Using Dynamic Averaging of Time Series Data

“…Finally, the output is calculated as the arithmetic mean of all individual tree predictions [26]. The random forest has many of advantageous aspects that make it very suitable for many prediction problems.…”

Jointly Learning Energy Expenditures and Activities Using Egocentric Multimodal Signals

“…Sensor-based Activity Recognition A number of works have used non-visual wearable sensors for activity recognition, including accelerometers and heart rate sensors [16,9,47]. Bao et al [7] use multiple accelerometers on the hip, wrist, arm, ankle and thigh to classify 20 classes of everyday household activities.…”

“…However, this method is highly inconvenient. As a proxy, measurements of acceleration and heart rate are widely used [39], e.g., multiple accelerometers [40,45,11], heart rate [10], indirect estimation from heart rate and oxygen uptake relationships [34], and heart rate in combination with accelerations [16]. However, estimating energy expenditure from visual data has not yet been explored.…”

Applied CyberGIS in the Age of Complex Spatial Health Data

“…Sitting below the workflow(s) are the systems and software deployments required for reliable execution, as well as the data communication layer for efficient data organization with is built upon the existing DELPHI infrastructure and adds in a PostgreSQL geodatabase for spatial data. A key element of SPACES is the creation of common data elements through existing physical activity and spatial energetics processing algorithms such as the Personal Activity Location Measurement System (Demchak et al 2012), machine learning algorithms (Ellis et al 2014), and spatial element extraction and analysis (Thierry et al 2013). Common data elements and outputs can directly enhance and influence the training efforts for the health research community and promote common data organization standards.…”