A hierarchical meta-classifier for 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

2 Citations (Scopus)

Abstract

This paper proposes a multi-level meta-classifier for identifying human activities based on accelerometer data. The training data consists of 77 subjects performing a combination of 23 different activities and monitored using a single hip-worn triaxial accelerometer. Time and frequency based features were extracted from two-second windows of raw accelerometer data and a subset of the features, together with demographic information, was selected for classification. The activities were divided into five activity groups: non-ambulatory activities, walking, running, climbing upstairs, and climbing downstairs. Multiple classification techniques were tested for each classifier level and groups. Random forests were found to perform comparatively better at each level. Based upon those tests, a 3-level hierarchical classifier, consisting of 5 random forest classifiers, was built. At the first level, the non-ambulatory activities are separated from the rest At the second, the ambulatory activities are divided into four activity groups. At the final level, the activities are classified individually. Accuracy on test sets was found to be approximately 87% overall for individual activities and 94% at the activity group level. These results compare favorably to contemporary results in classifying human activity.

Original languageEnglish (US)
Title of host publicationProceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages81-86
Number of pages6
ISBN (Electronic)9781509061662
DOIs
StatePublished - Jan 31 2017
Event15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016 - Anaheim, United States
Duration: Dec 18 2016Dec 20 2016

Other

Other15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016
CountryUnited States
CityAnaheim
Period12/18/1612/20/16

Fingerprint

Classifiers
Accelerometers

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Computer Science Applications

Cite this

Niazi, A. H., Yazdansepas, D., Gay, J. L., Maier, F. W., Ramaswamy, L., Rasheed, K., & Buman, M. (2017). A hierarchical meta-classifier for human activity recognition. In Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016 (pp. 81-86). [7838126] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICMLA.2016.16

A hierarchical meta-classifier for human activity recognition. / Niazi, Anzah H.; Yazdansepas, Delaram; Gay, Jennifer L.; Maier, Frederick W.; Ramaswamy, Lakshmish; Rasheed, Khaled; Buman, Matthew.

Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc., 2017. p. 81-86 7838126.

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

Niazi, AH, Yazdansepas, D, Gay, JL, Maier, FW, Ramaswamy, L, Rasheed, K & Buman, M 2017, A hierarchical meta-classifier for human activity recognition. in Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016., 7838126, Institute of Electrical and Electronics Engineers Inc., pp. 81-86, 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016, Anaheim, United States, 12/18/16. https://doi.org/10.1109/ICMLA.2016.16
Niazi AH, Yazdansepas D, Gay JL, Maier FW, Ramaswamy L, Rasheed K et al. A hierarchical meta-classifier for human activity recognition. In Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc. 2017. p. 81-86. 7838126 https://doi.org/10.1109/ICMLA.2016.16
Niazi, Anzah H. ; Yazdansepas, Delaram ; Gay, Jennifer L. ; Maier, Frederick W. ; Ramaswamy, Lakshmish ; Rasheed, Khaled ; Buman, Matthew. / A hierarchical meta-classifier for human activity recognition. Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 81-86
@inproceedings{cd696a726f9046c0b2eddbf1ce2084a7,
title = "A hierarchical meta-classifier for human activity recognition",
abstract = "This paper proposes a multi-level meta-classifier for identifying human activities based on accelerometer data. The training data consists of 77 subjects performing a combination of 23 different activities and monitored using a single hip-worn triaxial accelerometer. Time and frequency based features were extracted from two-second windows of raw accelerometer data and a subset of the features, together with demographic information, was selected for classification. The activities were divided into five activity groups: non-ambulatory activities, walking, running, climbing upstairs, and climbing downstairs. Multiple classification techniques were tested for each classifier level and groups. Random forests were found to perform comparatively better at each level. Based upon those tests, a 3-level hierarchical classifier, consisting of 5 random forest classifiers, was built. At the first level, the non-ambulatory activities are separated from the rest At the second, the ambulatory activities are divided into four activity groups. At the final level, the activities are classified individually. Accuracy on test sets was found to be approximately 87{\%} overall for individual activities and 94{\%} at the activity group level. These results compare favorably to contemporary results in classifying human activity.",
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 = "31",
doi = "10.1109/ICMLA.2016.16",
language = "English (US)",
pages = "81--86",
booktitle = "Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - A hierarchical meta-classifier for 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 - 2017/1/31

Y1 - 2017/1/31

N2 - This paper proposes a multi-level meta-classifier for identifying human activities based on accelerometer data. The training data consists of 77 subjects performing a combination of 23 different activities and monitored using a single hip-worn triaxial accelerometer. Time and frequency based features were extracted from two-second windows of raw accelerometer data and a subset of the features, together with demographic information, was selected for classification. The activities were divided into five activity groups: non-ambulatory activities, walking, running, climbing upstairs, and climbing downstairs. Multiple classification techniques were tested for each classifier level and groups. Random forests were found to perform comparatively better at each level. Based upon those tests, a 3-level hierarchical classifier, consisting of 5 random forest classifiers, was built. At the first level, the non-ambulatory activities are separated from the rest At the second, the ambulatory activities are divided into four activity groups. At the final level, the activities are classified individually. Accuracy on test sets was found to be approximately 87% overall for individual activities and 94% at the activity group level. These results compare favorably to contemporary results in classifying human activity.

AB - This paper proposes a multi-level meta-classifier for identifying human activities based on accelerometer data. The training data consists of 77 subjects performing a combination of 23 different activities and monitored using a single hip-worn triaxial accelerometer. Time and frequency based features were extracted from two-second windows of raw accelerometer data and a subset of the features, together with demographic information, was selected for classification. The activities were divided into five activity groups: non-ambulatory activities, walking, running, climbing upstairs, and climbing downstairs. Multiple classification techniques were tested for each classifier level and groups. Random forests were found to perform comparatively better at each level. Based upon those tests, a 3-level hierarchical classifier, consisting of 5 random forest classifiers, was built. At the first level, the non-ambulatory activities are separated from the rest At the second, the ambulatory activities are divided into four activity groups. At the final level, the activities are classified individually. Accuracy on test sets was found to be approximately 87% overall for individual activities and 94% at the activity group level. These results compare favorably to contemporary results in classifying human activity.

UR - http://www.scopus.com/inward/record.url?scp=85015446140&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015446140&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2016.16

DO - 10.1109/ICMLA.2016.16

M3 - Conference contribution

AN - SCOPUS:85015446140

SP - 81

EP - 86

BT - Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -