Feature Selection for Activity Recognition from Smartphone Accelerometer Data

Juan C. Quiroz, Amit Banerjee, Sergiu M. Dascalu, Sian Lun Lau

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

We use the public Human Activity Recognition Using Smartphones (HARUS) data-set to investigate and identify the most informative features for determining the physical activity performed by a user based on smartphone accelerometer and gyroscope data. The HARUS data-set includes 561 time domain and frequency domain features extracted from sensor readings collected from a smartphone carried by 30 users while performing specific activities. We compare the performance of a decision tree, support vector machines, Naive Bayes, multilayer perceptron, and bagging. We report the various classification performances of these algorithms for subject independent cases. Our results show that bagging and the multilayer perceptron achieve the highest classification accuracies across all feature sets. In addition, the signal from gravity contains the most information for classification of activities in the HARUS data-set.

Original languageEnglish (US)
Pages (from-to)1-9
Number of pages9
JournalIntelligent Automation and Soft Computing
DOIs
StateAccepted/In press - Jul 19 2017

Fingerprint

Activity Recognition
Smartphones
Accelerometer
Accelerometers
Feature Selection
Feature extraction
Bagging
Perceptron
Multilayer
Multilayer neural networks
Naive Bayes
Gyroscope
Decision tree
Frequency Domain
Gyroscopes
Time Domain
Support Vector Machine
Gravity
Decision trees
Support vector machines

All Science Journal Classification (ASJC) codes

  • Software
  • Theoretical Computer Science
  • Computational Theory and Mathematics
  • Artificial Intelligence

Cite this

@article{e7fd368a8d4e42ecae6a38bfbbd69b1a,
title = "Feature Selection for Activity Recognition from Smartphone Accelerometer Data",
abstract = "We use the public Human Activity Recognition Using Smartphones (HARUS) data-set to investigate and identify the most informative features for determining the physical activity performed by a user based on smartphone accelerometer and gyroscope data. The HARUS data-set includes 561 time domain and frequency domain features extracted from sensor readings collected from a smartphone carried by 30 users while performing specific activities. We compare the performance of a decision tree, support vector machines, Naive Bayes, multilayer perceptron, and bagging. We report the various classification performances of these algorithms for subject independent cases. Our results show that bagging and the multilayer perceptron achieve the highest classification accuracies across all feature sets. In addition, the signal from gravity contains the most information for classification of activities in the HARUS data-set.",
author = "Quiroz, {Juan C.} and Amit Banerjee and Dascalu, {Sergiu M.} and Lau, {Sian Lun}",
year = "2017",
month = "7",
day = "19",
doi = "10.1080/10798587.2017.1342400",
language = "English (US)",
pages = "1--9",
journal = "Intelligent Automation and Soft Computing",
issn = "1079-8587",
publisher = "AutoSoft Press",

}

Feature Selection for Activity Recognition from Smartphone Accelerometer Data. / Quiroz, Juan C.; Banerjee, Amit; Dascalu, Sergiu M.; Lau, Sian Lun.

In: Intelligent Automation and Soft Computing, 19.07.2017, p. 1-9.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Feature Selection for Activity Recognition from Smartphone Accelerometer Data

AU - Quiroz, Juan C.

AU - Banerjee, Amit

AU - Dascalu, Sergiu M.

AU - Lau, Sian Lun

PY - 2017/7/19

Y1 - 2017/7/19

N2 - We use the public Human Activity Recognition Using Smartphones (HARUS) data-set to investigate and identify the most informative features for determining the physical activity performed by a user based on smartphone accelerometer and gyroscope data. The HARUS data-set includes 561 time domain and frequency domain features extracted from sensor readings collected from a smartphone carried by 30 users while performing specific activities. We compare the performance of a decision tree, support vector machines, Naive Bayes, multilayer perceptron, and bagging. We report the various classification performances of these algorithms for subject independent cases. Our results show that bagging and the multilayer perceptron achieve the highest classification accuracies across all feature sets. In addition, the signal from gravity contains the most information for classification of activities in the HARUS data-set.

AB - We use the public Human Activity Recognition Using Smartphones (HARUS) data-set to investigate and identify the most informative features for determining the physical activity performed by a user based on smartphone accelerometer and gyroscope data. The HARUS data-set includes 561 time domain and frequency domain features extracted from sensor readings collected from a smartphone carried by 30 users while performing specific activities. We compare the performance of a decision tree, support vector machines, Naive Bayes, multilayer perceptron, and bagging. We report the various classification performances of these algorithms for subject independent cases. Our results show that bagging and the multilayer perceptron achieve the highest classification accuracies across all feature sets. In addition, the signal from gravity contains the most information for classification of activities in the HARUS data-set.

UR - http://www.scopus.com/inward/record.url?scp=85024478582&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85024478582&partnerID=8YFLogxK

U2 - 10.1080/10798587.2017.1342400

DO - 10.1080/10798587.2017.1342400

M3 - Article

AN - SCOPUS:85024478582

SP - 1

EP - 9

JO - Intelligent Automation and Soft Computing

JF - Intelligent Automation and Soft Computing

SN - 1079-8587

ER -