Statistical analysis of window sizes and sampling rates in human activity recognition

Anzah H. Niazi, Delaram Yazdansepas, Jennifer L. Gay, Frederick W. Maier, Lakshmish Ramaswamy, Khaled Rasheed, Matthew Buman

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

Accelerometers are the most common device for data collection in the field of Human Activity Recognition (HAR). This data is recorded at a particular sampling rate and then usually separated into time windows before classification takes place. Though the sampling rate and window size can have a significant impact on the accuracy of the trained classifier, there has been relatively little research on their role in activity recognition. This paper presents a statistical analysis on the effect the sampling rate and window sizes on HAR data classification. The raw data used in the analysis was collected from a hip-worn Actigraphy G3X+ at 100Hz from 77 subjects performing 23 different activities. It was then re-sampled and divided into windows of varying sizes and trained using a single data classifier. A weighted least squares linear regression model was developed and two-way factorial ANOVA was used to analyze the effects of sampling rate and window size for different activity types and demographic categories. Based upon this analysis, we find that 10-second windows recorded at 50Hz perform statistically better than other combinations of window size and sampling rate.

Original languageEnglish (US)
Title of host publicationHEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017
EditorsHugo Gamboa, Ana Fred, Egon L. van den Broek, Mario Vaz
PublisherSciTePress
Pages319-325
Number of pages7
Volume5
ISBN (Electronic)9789897582134
StatePublished - Jan 1 2017
Event10th International Conference on Health Informatics, HEALTHINF 2017 - Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017 - Porto, Portugal
Duration: Feb 21 2017Feb 23 2017

Other

Other10th International Conference on Health Informatics, HEALTHINF 2017 - Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017
CountryPortugal
CityPorto
Period2/21/172/23/17

Fingerprint

Human Activities
Linear Models
Statistical methods
Actigraphy
Sampling
Least-Squares Analysis
Hip
Analysis of Variance
Demography
Classifiers
Equipment and Supplies
Research
Analysis of variance (ANOVA)
Accelerometers
Linear regression

Keywords

  • Actigraph g3x+
  • Analysis of Variance
  • Body-worn Accelerometers
  • Data Mining
  • Human Activity Recognition
  • Random Forests
  • Sampling Rate
  • Weighted Least Squares
  • WEKA
  • Window Size

ASJC Scopus subject areas

  • Biomedical Engineering
  • Electrical and Electronic Engineering
  • Health Informatics
  • Health Information Management

Cite this

Niazi, A. H., Yazdansepas, D., Gay, J. L., Maier, F. W., Ramaswamy, L., Rasheed, K., & Buman, M. (2017). Statistical analysis of window sizes and sampling rates in human activity recognition. In H. Gamboa, A. Fred, E. L. van den Broek, & M. Vaz (Eds.), HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017 (Vol. 5, pp. 319-325). SciTePress.

Statistical analysis of window sizes and sampling rates in human activity recognition. / Niazi, Anzah H.; Yazdansepas, Delaram; Gay, Jennifer L.; Maier, Frederick W.; Ramaswamy, Lakshmish; Rasheed, Khaled; Buman, Matthew.

HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017. ed. / Hugo Gamboa; Ana Fred; Egon L. van den Broek; Mario Vaz. Vol. 5 SciTePress, 2017. p. 319-325.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Niazi, AH, Yazdansepas, D, Gay, JL, Maier, FW, Ramaswamy, L, Rasheed, K & Buman, M 2017, Statistical analysis of window sizes and sampling rates in human activity recognition. in H Gamboa, A Fred, EL van den Broek & M Vaz (eds), HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017. vol. 5, SciTePress, pp. 319-325, 10th International Conference on Health Informatics, HEALTHINF 2017 - Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017, Porto, Portugal, 2/21/17.
Niazi AH, Yazdansepas D, Gay JL, Maier FW, Ramaswamy L, Rasheed K et al. Statistical analysis of window sizes and sampling rates in human activity recognition. In Gamboa H, Fred A, van den Broek EL, Vaz M, editors, HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017. Vol. 5. SciTePress. 2017. p. 319-325
Niazi, Anzah H. ; Yazdansepas, Delaram ; Gay, Jennifer L. ; Maier, Frederick W. ; Ramaswamy, Lakshmish ; Rasheed, Khaled ; Buman, Matthew. / Statistical analysis of window sizes and sampling rates in human activity recognition. HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017. editor / Hugo Gamboa ; Ana Fred ; Egon L. van den Broek ; Mario Vaz. Vol. 5 SciTePress, 2017. pp. 319-325
@inproceedings{314f9598718b4db8be1e21d8e9624075,
title = "Statistical analysis of window sizes and sampling rates in human activity recognition",
abstract = "Accelerometers are the most common device for data collection in the field of Human Activity Recognition (HAR). This data is recorded at a particular sampling rate and then usually separated into time windows before classification takes place. Though the sampling rate and window size can have a significant impact on the accuracy of the trained classifier, there has been relatively little research on their role in activity recognition. This paper presents a statistical analysis on the effect the sampling rate and window sizes on HAR data classification. The raw data used in the analysis was collected from a hip-worn Actigraphy G3X+ at 100Hz from 77 subjects performing 23 different activities. It was then re-sampled and divided into windows of varying sizes and trained using a single data classifier. A weighted least squares linear regression model was developed and two-way factorial ANOVA was used to analyze the effects of sampling rate and window size for different activity types and demographic categories. Based upon this analysis, we find that 10-second windows recorded at 50Hz perform statistically better than other combinations of window size and sampling rate.",
keywords = "Actigraph g3x+, Analysis of Variance, Body-worn Accelerometers, Data Mining, Human Activity Recognition, Random Forests, Sampling Rate, Weighted Least Squares, WEKA, Window Size",
author = "Niazi, {Anzah H.} and Delaram Yazdansepas and Gay, {Jennifer L.} and Maier, {Frederick W.} and Lakshmish Ramaswamy and Khaled Rasheed and Matthew Buman",
year = "2017",
month = "1",
day = "1",
language = "English (US)",
volume = "5",
pages = "319--325",
editor = "Hugo Gamboa and Ana Fred and {van den Broek}, {Egon L.} and Mario Vaz",
booktitle = "HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017",
publisher = "SciTePress",

}

TY - GEN

T1 - Statistical analysis of window sizes and sampling rates in human activity recognition

AU - Niazi, Anzah H.

AU - Yazdansepas, Delaram

AU - Gay, Jennifer L.

AU - Maier, Frederick W.

AU - Ramaswamy, Lakshmish

AU - Rasheed, Khaled

AU - Buman, Matthew

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Accelerometers are the most common device for data collection in the field of Human Activity Recognition (HAR). This data is recorded at a particular sampling rate and then usually separated into time windows before classification takes place. Though the sampling rate and window size can have a significant impact on the accuracy of the trained classifier, there has been relatively little research on their role in activity recognition. This paper presents a statistical analysis on the effect the sampling rate and window sizes on HAR data classification. The raw data used in the analysis was collected from a hip-worn Actigraphy G3X+ at 100Hz from 77 subjects performing 23 different activities. It was then re-sampled and divided into windows of varying sizes and trained using a single data classifier. A weighted least squares linear regression model was developed and two-way factorial ANOVA was used to analyze the effects of sampling rate and window size for different activity types and demographic categories. Based upon this analysis, we find that 10-second windows recorded at 50Hz perform statistically better than other combinations of window size and sampling rate.

AB - Accelerometers are the most common device for data collection in the field of Human Activity Recognition (HAR). This data is recorded at a particular sampling rate and then usually separated into time windows before classification takes place. Though the sampling rate and window size can have a significant impact on the accuracy of the trained classifier, there has been relatively little research on their role in activity recognition. This paper presents a statistical analysis on the effect the sampling rate and window sizes on HAR data classification. The raw data used in the analysis was collected from a hip-worn Actigraphy G3X+ at 100Hz from 77 subjects performing 23 different activities. It was then re-sampled and divided into windows of varying sizes and trained using a single data classifier. A weighted least squares linear regression model was developed and two-way factorial ANOVA was used to analyze the effects of sampling rate and window size for different activity types and demographic categories. Based upon this analysis, we find that 10-second windows recorded at 50Hz perform statistically better than other combinations of window size and sampling rate.

KW - Actigraph g3x+

KW - Analysis of Variance

KW - Body-worn Accelerometers

KW - Data Mining

KW - Human Activity Recognition

KW - Random Forests

KW - Sampling Rate

KW - Weighted Least Squares

KW - WEKA

KW - Window Size

UR - http://www.scopus.com/inward/record.url?scp=85051750855&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85051750855&partnerID=8YFLogxK

M3 - Conference contribution

VL - 5

SP - 319

EP - 325

BT - HEALTHINF 2017 - 10th International Conference on Health Informatics, Proceedings; Part of 10th International Joint Conference on Biomedical Engineering Systems and Technologies, BIOSTEC 2017

A2 - Gamboa, Hugo

A2 - Fred, Ana

A2 - van den Broek, Egon L.

A2 - Vaz, Mario

PB - SciTePress

ER -