Comparing Supervised Learning Techniques on the Task of Physical Activity Recognition

The objective of this study was to compare the performance of base-level and meta-level classifiers on the task of physical activity recognition. Five wireless kinematic sensors were attached to each subject (n = 25) while they completed a range of basic physical activities in a controlled laboratory setting. Subjects were then asked to carry out similar self-annotated physical activities in a random order and in an unsupervised environment. A combination of time-domain and frequency-domain features was extracted from the sensor data including the first four central moments, zero-crossing rate, average magnitude, sensor cross correlation, sensor autocorrelation, spectral entropy, and dominant frequency components. A reduced feature set was generated using a wrapper subset evaluation technique with a linear forward search and this feature set was employed for classifier comparison. The meta-level classifier AdaBoostM1 with C4.5 Graft as its base-level classifier achieved an overall accuracy of 95%. Equal sized datasets of subject-independent data and subject-dependent data were used to train this classifier and high recognition rates could be achieved without the need for user specific training. Furthermore, it was found that an accuracy of 88% could be achieved using data from the ankle and wrist sensors only.

[1]  Nigel H. Lovell,et al.  Implementation of a real-time human movement classifier using a triaxial accelerometer for ambulatory monitoring , 2006, IEEE Transactions on Information Technology in Biomedicine.

[2]  Fearghal Morgan,et al.  A preliminary study of using wireless kinematic sensors to identify basic Activities of Daily Living , 2008, 2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[3]  Usdhhs.  Physical Activity and Health: A Report of the Surgeon General , 1996 .

[4]  E. Sobngwi,et al.  Physical activity and its relationship with obesity, hypertension and diabetes in urban and rural Cameroon , 2002, International Journal of Obesity.

[5]  Ian H. Witten,et al.  Data mining: practical machine learning tools and techniques with Java implementations , 2002, SGMD.

[6]  Gary M. Weiss,et al.  Activity recognition using cell phone accelerometers , 2011, SKDD.

[7]  Daniel P. Siewiorek,et al.  Activity recognition and monitoring using multiple sensors on different body positions , 2006, International Workshop on Wearable and Implantable Body Sensor Networks (BSN'06).

[8]  P Shekelle,et al.  Differences between Patients' and Physicians' Evaluations of Outcome after Total Hip Arthroplasty* , 1996, The Journal of bone and joint surgery. American volume.

[9]  Tae-Seong Kim,et al.  A Triaxial Accelerometer-Based Physical-Activity Recognition via Augmented-Signal Features and a Hierarchical Recognizer , 2010, IEEE Transactions on Information Technology in Biomedicine.

[10]  John W. Sammon,et al.  A Nonlinear Mapping for Data Structure Analysis , 1969, IEEE Transactions on Computers.

[11]  Michael L. Littman,et al.  Activity Recognition from Accelerometer Data , 2005, AAAI.

[12]  Bernt Schiele,et al.  Analyzing features for activity recognition , 2005, sOc-EUSAI '05.

[13]  Angelo M. Sabatini,et al.  Machine Learning Methods for Classifying Human Physical Activity from On-Body Accelerometers , 2010, Sensors.

[14]  Ling Bao,et al.  Activity Recognition from User-Annotated Acceleration Data , 2004, Pervasive.

[15]  K R Westerterp,et al.  Physical activity and parameters of aging: a physiological perspective. , 2001, The journals of gerontology. Series A, Biological sciences and medical sciences.

[16]  J. D. Janssen,et al.  A triaxial accelerometer and portable data processing unit for the assessment of daily physical activity , 1997, IEEE Transactions on Biomedical Engineering.

[17]  Chih-Jen Lin,et al.  A Practical Guide to Support Vector Classication , 2008 .

[18]  Guang-Zhong Yang,et al.  Real-Time Pervasive Monitoring for Postoperative Care , 2007, BSN.

[19]  Daniel Olgu ´ õn,et al.  Human Activity Recognition: Accuracy across Common Locations for Wearable Sensors , 2006 .

[20]  C. N. Scanaill,et al.  Long-term telemonitoring of mobility trends of elderly people using SMS messaging , 2006, IEEE Transactions on Information Technology in Biomedicine.

[21]  Luca Palmerini,et al.  Feature Selection for Accelerometer-Based Posture Analysis in Parkinson's Disease , 2011, IEEE Transactions on Information Technology in Biomedicine.

[22]  Jorunn L Helbostad,et al.  Estimation of gait cycle characteristics by trunk accelerometry. , 2004, Journal of biomechanics.

[23]  Chih-Jen Lin,et al.  LIBSVM: A library for support vector machines , 2011, TIST.

[24]  Takuya Maekawa,et al.  Object-Based Activity Recognition with Heterogeneous Sensors on Wrist , 2010, Pervasive.

[25]  G M Lyons,et al.  A description of an accelerometer-based mobility monitoring technique. , 2005, Medical engineering & physics.

[26]  Kay Römer,et al.  Practical time synchronization for Bluetooth Scatternets , 2007, 2007 Fourth International Conference on Broadband Communications, Networks and Systems (BROADNETS '07).

[27]  Arie Ben-David,et al.  Comparison of classification accuracy using Cohen's Weighted Kappa , 2008, Expert Syst. Appl..

[28]  C J Snijders,et al.  Computerized analysis of daily life motor activity for ambulatory monitoring. , 1997, Technology and health care : official journal of the European Society for Engineering and Medicine.

[29]  Bernt Schiele,et al.  Towards Less Supervision in Activity Recognition from Wearable Sensors , 2006, 2006 10th IEEE International Symposium on Wearable Computers.

[30]  Jiawei Han,et al.  Data Mining: Concepts and Techniques , 2000 .

[31]  G. ÓLaighin,et al.  Direct measurement of human movement by accelerometry. , 2008, Medical engineering & physics.

[32]  A. Manley.  Physical Activity And Health: A Report Of The Surgeon General , 2004 .

[33]  K. Fox.  The influence of physical activity on mental well-being. , 1999, Public health nutrition.

[34]  Tatsuo Nakajima,et al.  Feature Selection and Activity Recognition from Wearable Sensors , 2006, UCS.

[35]  A. Bourke,et al.  Distinguishing falls from normal ADL using vertical velocity profiles , 2007, 2007 29th Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[36]  Gaetano Borriello,et al.  A Practical Approach to Recognizing Physical Activities , 2006, Pervasive.

[37]  Victor R. Preedy,et al.  International Classification of Impairment, Disabilities and Handicaps , 2010 .

[38]  A M Prentice,et al.  Human energy expenditure in affluent societies: an analysis of 574 doubly-labelled water measurements. , 1996, European journal of clinical nutrition.