Statistical Learning Methods to Predict Activity Intensity from Body-Worn Accelerometers

Physical activity, especially when performed at moderate or vigorous intensity, has short- and long-term health benefits, but measurement of free-living physical activity is challenging. Accelerometers are popular tools to assess physical activity, although accuracy of conventional accelerometer analysis methods is suboptimal. This study developed and tested statistical learning models for assessing activity intensity from body-worn accelerometers. Twenty-eight adults performed 10-21 activities of daily living in two visits while wearing four accelerometers (right hip, right ankle, both wrists). Accelerometer placement is of crucial practical concern and this paper addresses this issue. Boosting, bagging, random forest and decision tree models were created for each accelerometer and for two-, three-, and four-accelerometer combinations to predict activity intensity. Research staff observations of activity intensity served as the criterion. Point estimates of error for the ankle accelerometer were 2.2-4.7 percentage points lower than other single-accelerometer placements, and the left wrist-ankle combination had errors 0.8-5.8 percentage points lower than other two-accelerometer combinations. Decision trees had poorer accuracy than the other models. Using an accelerometer worn on the lower limb, by itself or in combination with an upper-limb accelerometer, appears to offer optimal accuracy for activity intensity measurement.

[1]  Shrikant I Bangdiwala Graphical aids for visualizing and interpreting patterns in departures from agreement in ordinal categorical observer agreement data , 2017, Journal of biopharmaceutical statistics.

[2]  W. Loh,et al.  A regression tree approach to identifying subgroups with differential treatment effects , 2014, Statistics in medicine.

[3]  Dinesh John,et al.  ActiGraph and Actical physical activity monitors: a peek under the hood. , 2012, Medicine and science in sports and exercise.

[4]  Ke Lu,et al.  Fusion of Heart Rate, Respiration and Motion Measurements from a Wearable Sensor System to Enhance Energy Expenditure Estimation , 2018, Sensors.

[5]  DAVID G. KENDALL,et al.  Introduction to Mathematical Statistics , 1947, Nature.

[6]  J. Staudenmayer,et al.  Methods to estimate aspects of physical activity and sedentary behavior from high-frequency wrist accelerometer measurements. , 2015, Journal of applied physiology.

[7]  Maarit Kangas,et al.  Calibration and validation of accelerometer-based activity monitors: A systematic review of machine-learning approaches. , 2019, Gait & posture.

[8]  P S Freedson,et al.  Calibration of the Computer Science and Applications, Inc. accelerometer. , 1998, Medicine and science in sports and exercise.

[9]  M. Guazzi,et al.  Assessing Physical Activity as a Core Component in Cardiac Rehabilitation: A POSITION STATEMENT OF THE AMERICAN ASSOCIATION OF CARDIOVASCULAR AND PULMONARY REHABILITATION , 2016, Journal of cardiopulmonary rehabilitation and prevention.

[10]  K. Patrick,et al.  Physical Activity and Public Health: A Recommendation From the Centers for Disease Control and Prevention and the American College of Sports Medicine , 1995 .

[11]  John Staudenmayer,et al.  Direct observation is a valid criterion for estimating physical activity and sedentary behavior. , 2014, Journal of physical activity & health.

[12]  Yoav Freund,et al.  A decision-theoretic generalization of on-line learning and an application to boosting , 1995, EuroCOLT.

[13]  Karin A Pfeiffer,et al.  Energy Expenditure Prediction Using Raw Accelerometer Data in Simulated Free Living. , 2015, Medicine and science in sports and exercise.

[14]  W. Loh,et al.  REGRESSION TREES WITH UNBIASED VARIABLE SELECTION AND INTERACTION DETECTION , 2002 .

[15]  Wei-Yin Loh,et al.  Classification and regression trees , 2011, WIREs Data Mining Knowl. Discov..

[16]  Dinesh John,et al.  Performance of Activity Classification Algorithms in Free-Living Older Adults. , 2016, Medicine and science in sports and exercise.

[17]  Laura D. Ellingson,et al.  Sedentary Behavior Research Network (SBRN) – Terminology Consensus Project process and outcome , 2017, International Journal of Behavioral Nutrition and Physical Activity.

[18]  Richard P Troiano,et al.  Evolution of accelerometer methods for physical activity research , 2014, British Journal of Sports Medicine.

[19]  Bo Dong,et al.  Energy-aware activity classification using wearable sensor networks , 2013, Defense, Security, and Sensing.

[20]  P. Freedson,et al.  Amount of time spent in sedentary behaviors in the United States, 2003-2004. , 2008, American journal of epidemiology.

[21]  Morten Wang Fagerland,et al.  Does physical activity attenuate, or even eliminate, the detrimental association of sitting time with mortality? A harmonised meta-analysis of data from more than 1 million men and women , 2016, The Lancet.

[22]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[23]  E. Forsum,et al.  Evaluations of Actiheart, IDEEA® and RT3 monitors for estimating activity energy expenditure in free-living women , 2013, Journal of Nutritional Science.

[24]  Scott E Crouter,et al.  Estimating Energy Expenditure with ActiGraph GT9X Inertial Measurement Unit , 2017, Medicine and science in sports and exercise.

[25]  M. Black,et al.  Ankle Accelerometry for Assessing Physical Activity Among Adolescent Girls: Threshold Determination, Validity, Reliability, and Feasibility , 2015, Research quarterly for exercise and sport.

[26]  Lindsay P. Toth,et al.  Video-Recorded Validation of Wearable Step Counters under Free-living Conditions , 2018, Medicine and science in sports and exercise.

[27]  Raymond C Browning,et al.  A comparison of energy expenditure estimation of several physical activity monitors. , 2013, Medicine and science in sports and exercise.

[28]  Gert R. G. Lanckriet,et al.  Objective Assessment of Physical Activity: Classifiers for Public Health. , 2016, Medicine and science in sports and exercise.

[29]  Vinod Chandran,et al.  Physical Activity Recognition Using Posterior-Adapted Class-Based Fusion of Multiaccelerometer Data , 2017, IEEE Journal of Biomedical and Health Informatics.

[30]  James M. Pivarnik,et al.  Validation and Comparison of Accelerometers Worn on the Hip, Thigh, and Wrists for Measuring Physical Activity and Sedentary Behavior , 2016, AIMS public health.

[31]  K A Mackintosh,et al.  Investigating optimal accelerometer placement for energy expenditure prediction in children using a machine learning approach , 2016, Physiological measurement.

[32]  Bradford S. Westgate,et al.  Cross-validation and out-of-sample testing of physical activity intensity predictions with a wrist-worn accelerometer. , 2018, Journal of applied physiology.

[33]  I. Lipkovich,et al.  Subgroup identification based on differential effect search—A recursive partitioning method for establishing response to treatment in patient subpopulations , 2011, Statistics in medicine.

[34]  Munni Begum,et al.  Comparison of linear and non-linear models for predicting energy expenditure from raw accelerometer data , 2017, Physiological measurement.

[35]  J. R. Landis,et al.  The measurement of observer agreement for categorical data. , 1977, Biometrics.

[36]  S. Carlson,et al.  The Physical Activity Guidelines for Americans , 2018, JAMA.

[37]  Jacob Cohen,et al.  Weighted kappa: Nominal scale agreement provision for scaled disagreement or partial credit. , 1968 .

[38]  David R Bassett,et al.  2011 Compendium of Physical Activities: a second update of codes and MET values. , 2011, Medicine and science in sports and exercise.

[39]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[40]  Ruairi O'Driscoll,et al.  How well do activity monitors estimate energy expenditure? A systematic review and meta-analysis of the validity of current technologies , 2020, British Journal of Sports Medicine.

[41]  Leonard A Kaminsky,et al.  Estimating relative intensity using individualized accelerometer cutpoints: the importance of fitness level , 2013, BMC Medical Research Methodology.

[42]  Greg J. Petrucci,et al.  Estimating Sedentary Time from a Hip- and Wrist-worn Accelerometer. , 2019, Medicine and science in sports and exercise.

[43]  Leonard A Kaminsky,et al.  Variability of Objectively Measured Sedentary Behavior. , 2016, Medicine and science in sports and exercise.

[44]  Hyunjoong Kim,et al.  Classification Trees With Unbiased Multiway Splits , 2001 .