Optimal model selection for posture recognition in home-based healthcare

Shumei Zhang, PJ McCullagh, CD Nugent, H Zheng, Matthias Baumgarten

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

This paper investigates optimal model selectionfor posture recognition. Accuracy and computational timeare related to the trained model in a supervised classification.An optimal model selection is important for a reliableactivity monitoring system. Conventional guidance onmodel training uses large instances of randomly selecteddata in order to characterize the classes. A new approach tothe training of a multiclass support vector machine (SVM)model suited to limited training sets such as used in posturerecognition is provided. This approach picks a smalltraining set from misclassified data to improve an initialmodel in an iterative and incremental fashion. In addition,a two step grid-search algorithm is used for the parameterssetting. The best parameters were chosen according to thetesting accuracy rather than conventional validating accuracy.This new approach for model selection was evaluatedagainst conventional approaches in an activity classificationstudy. Nine everyday postures were classified from abelt-worn smart phone’s accelerometer data. The classificationderived from the small training set and the conventionalrandomly selected training set differed in twoaspects: classification performance to new data (85.1%Pick-out small training set vs. 70.3% conventional largetraining set) and computational efficiency (improved 28%).
LanguageEnglish
Pages1-14
JournalInternational Journal of Machine Learning and Cybernetics
Volume2
Issue number1
DOIs
Publication statusPublished - Mar 2011

Fingerprint

Computational efficiency
Accelerometers
Support vector machines
Monitoring

Keywords

  • Optimal model
  • Posture recognition
  • Accelerometer
  • Multi-class SVM

Cite this

@article{0b4b5e167936480db9de94f2abf57df4,
title = "Optimal model selection for posture recognition in home-based healthcare",
abstract = "This paper investigates optimal model selectionfor posture recognition. Accuracy and computational timeare related to the trained model in a supervised classification.An optimal model selection is important for a reliableactivity monitoring system. Conventional guidance onmodel training uses large instances of randomly selecteddata in order to characterize the classes. A new approach tothe training of a multiclass support vector machine (SVM)model suited to limited training sets such as used in posturerecognition is provided. This approach picks a smalltraining set from misclassified data to improve an initialmodel in an iterative and incremental fashion. In addition,a two step grid-search algorithm is used for the parameterssetting. The best parameters were chosen according to thetesting accuracy rather than conventional validating accuracy.This new approach for model selection was evaluatedagainst conventional approaches in an activity classificationstudy. Nine everyday postures were classified from abelt-worn smart phone’s accelerometer data. The classificationderived from the small training set and the conventionalrandomly selected training set differed in twoaspects: classification performance to new data (85.1{\%}Pick-out small training set vs. 70.3{\%} conventional largetraining set) and computational efficiency (improved 28{\%}).",
keywords = "Optimal model, Posture recognition, Accelerometer, Multi-class SVM",
author = "Shumei Zhang and PJ McCullagh and CD Nugent and H Zheng and Matthias Baumgarten",
year = "2011",
month = "3",
doi = "10.1007/s13042-010-0009-5",
language = "English",
volume = "2",
pages = "1--14",
journal = "International Journal of Machine Learning and Cybernetics",
issn = "1868-8071",
number = "1",

}

Optimal model selection for posture recognition in home-based healthcare. / Zhang, Shumei; McCullagh, PJ; Nugent, CD; Zheng, H; Baumgarten, Matthias.

In: International Journal of Machine Learning and Cybernetics, Vol. 2, No. 1, 03.2011, p. 1-14.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Optimal model selection for posture recognition in home-based healthcare

AU - Zhang, Shumei

AU - McCullagh, PJ

AU - Nugent, CD

AU - Zheng, H

AU - Baumgarten, Matthias

PY - 2011/3

Y1 - 2011/3

N2 - This paper investigates optimal model selectionfor posture recognition. Accuracy and computational timeare related to the trained model in a supervised classification.An optimal model selection is important for a reliableactivity monitoring system. Conventional guidance onmodel training uses large instances of randomly selecteddata in order to characterize the classes. A new approach tothe training of a multiclass support vector machine (SVM)model suited to limited training sets such as used in posturerecognition is provided. This approach picks a smalltraining set from misclassified data to improve an initialmodel in an iterative and incremental fashion. In addition,a two step grid-search algorithm is used for the parameterssetting. The best parameters were chosen according to thetesting accuracy rather than conventional validating accuracy.This new approach for model selection was evaluatedagainst conventional approaches in an activity classificationstudy. Nine everyday postures were classified from abelt-worn smart phone’s accelerometer data. The classificationderived from the small training set and the conventionalrandomly selected training set differed in twoaspects: classification performance to new data (85.1%Pick-out small training set vs. 70.3% conventional largetraining set) and computational efficiency (improved 28%).

AB - This paper investigates optimal model selectionfor posture recognition. Accuracy and computational timeare related to the trained model in a supervised classification.An optimal model selection is important for a reliableactivity monitoring system. Conventional guidance onmodel training uses large instances of randomly selecteddata in order to characterize the classes. A new approach tothe training of a multiclass support vector machine (SVM)model suited to limited training sets such as used in posturerecognition is provided. This approach picks a smalltraining set from misclassified data to improve an initialmodel in an iterative and incremental fashion. In addition,a two step grid-search algorithm is used for the parameterssetting. The best parameters were chosen according to thetesting accuracy rather than conventional validating accuracy.This new approach for model selection was evaluatedagainst conventional approaches in an activity classificationstudy. Nine everyday postures were classified from abelt-worn smart phone’s accelerometer data. The classificationderived from the small training set and the conventionalrandomly selected training set differed in twoaspects: classification performance to new data (85.1%Pick-out small training set vs. 70.3% conventional largetraining set) and computational efficiency (improved 28%).

KW - Optimal model

KW - Posture recognition

KW - Accelerometer

KW - Multi-class SVM

UR - http://www.springerlink.com

U2 - 10.1007/s13042-010-0009-5

DO - 10.1007/s13042-010-0009-5

M3 - Article

VL - 2

SP - 1

EP - 14

JO - International Journal of Machine Learning and Cybernetics

T2 - International Journal of Machine Learning and Cybernetics

JF - International Journal of Machine Learning and Cybernetics

SN - 1868-8071

IS - 1

ER -