Abstract

With the recent interest in physical therapy through sufficient physical activity, considerable efforts have been made to monitor and classify daily human activities, especially for people who need physical rehabilitation. In our previous study, we designed a classifier to identify 25 unique physical activities performed by 92 healthy participants between the ages of 20 and 65. In this study, with the use of a GENEActiv accelerometer to monitor a wide range of daily activities, we present a learning approach to identify unique activities performed by a varied group of participants with various health conditions. The dataset is comprised of 99 senior participants and 23 participants who are significantly taller in height than the general population, performing 8 unique activities. We have extracted 130 different features in time and frequency domain and selected the most efficient features with the sequential forward selection algorithm. With two stages of classification, the first is utilized for combining similar classes, and the second determines the final decision. We have tested two classifiers for our learning approach, the Gaussian mixture models (GMMs) and the hidden Markov models (HMMs) and compared their performances. We have improved the GMM classifier from our previous study and it has shown more promising results for this dataset. We achieved an accuracy of 88.92% when classifying the 8 unique activities with GMM and 93.5% with HMM when classifying 7 activities.

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.
Pages339-344
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

Hidden Markov models
Accelerometers
Classifiers
Sensors
Physical therapy
Patient rehabilitation
Health

ASJC Scopus subject areas

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

Cite this

Dutta, A., Ma, O., Toledo, M., Buman, M., & Bliss, D. (2017). Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data. In Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016 (pp. 339-344). [7838166] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICMLA.2016.64

Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data. / Dutta, Arindam; Ma, Owen; Toledo, Meynard; Buman, Matthew; Bliss, Daniel.

Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc., 2017. p. 339-344 7838166.

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

Dutta, A, Ma, O, Toledo, M, Buman, M & Bliss, D 2017, Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data. in Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016., 7838166, Institute of Electrical and Electronics Engineers Inc., pp. 339-344, 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016, Anaheim, United States, 12/18/16. https://doi.org/10.1109/ICMLA.2016.64
Dutta A, Ma O, Toledo M, Buman M, Bliss D. Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data. In Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc. 2017. p. 339-344. 7838166 https://doi.org/10.1109/ICMLA.2016.64
Dutta, Arindam ; Ma, Owen ; Toledo, Meynard ; Buman, Matthew ; Bliss, Daniel. / Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data. Proceedings - 2016 15th IEEE International Conference on Machine Learning and Applications, ICMLA 2016. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 339-344
@inproceedings{329cd6e085f14ef292d0a7c4a33bc022,
title = "Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data",
abstract = "With the recent interest in physical therapy through sufficient physical activity, considerable efforts have been made to monitor and classify daily human activities, especially for people who need physical rehabilitation. In our previous study, we designed a classifier to identify 25 unique physical activities performed by 92 healthy participants between the ages of 20 and 65. In this study, with the use of a GENEActiv accelerometer to monitor a wide range of daily activities, we present a learning approach to identify unique activities performed by a varied group of participants with various health conditions. The dataset is comprised of 99 senior participants and 23 participants who are significantly taller in height than the general population, performing 8 unique activities. We have extracted 130 different features in time and frequency domain and selected the most efficient features with the sequential forward selection algorithm. With two stages of classification, the first is utilized for combining similar classes, and the second determines the final decision. We have tested two classifiers for our learning approach, the Gaussian mixture models (GMMs) and the hidden Markov models (HMMs) and compared their performances. We have improved the GMM classifier from our previous study and it has shown more promising results for this dataset. We achieved an accuracy of 88.92{\%} when classifying the 8 unique activities with GMM and 93.5{\%} with HMM when classifying 7 activities.",
author = "Arindam Dutta and Owen Ma and Meynard Toledo and Matthew Buman and Daniel Bliss",
year = "2017",
month = "1",
day = "31",
doi = "10.1109/ICMLA.2016.64",
language = "English (US)",
pages = "339--344",
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 - Comparing Gaussian mixture model and hidden Markov model to classify unique physical activities from accelerometer sensor data

AU - Dutta, Arindam

AU - Ma, Owen

AU - Toledo, Meynard

AU - Buman, Matthew

AU - Bliss, Daniel

PY - 2017/1/31

Y1 - 2017/1/31

N2 - With the recent interest in physical therapy through sufficient physical activity, considerable efforts have been made to monitor and classify daily human activities, especially for people who need physical rehabilitation. In our previous study, we designed a classifier to identify 25 unique physical activities performed by 92 healthy participants between the ages of 20 and 65. In this study, with the use of a GENEActiv accelerometer to monitor a wide range of daily activities, we present a learning approach to identify unique activities performed by a varied group of participants with various health conditions. The dataset is comprised of 99 senior participants and 23 participants who are significantly taller in height than the general population, performing 8 unique activities. We have extracted 130 different features in time and frequency domain and selected the most efficient features with the sequential forward selection algorithm. With two stages of classification, the first is utilized for combining similar classes, and the second determines the final decision. We have tested two classifiers for our learning approach, the Gaussian mixture models (GMMs) and the hidden Markov models (HMMs) and compared their performances. We have improved the GMM classifier from our previous study and it has shown more promising results for this dataset. We achieved an accuracy of 88.92% when classifying the 8 unique activities with GMM and 93.5% with HMM when classifying 7 activities.

AB - With the recent interest in physical therapy through sufficient physical activity, considerable efforts have been made to monitor and classify daily human activities, especially for people who need physical rehabilitation. In our previous study, we designed a classifier to identify 25 unique physical activities performed by 92 healthy participants between the ages of 20 and 65. In this study, with the use of a GENEActiv accelerometer to monitor a wide range of daily activities, we present a learning approach to identify unique activities performed by a varied group of participants with various health conditions. The dataset is comprised of 99 senior participants and 23 participants who are significantly taller in height than the general population, performing 8 unique activities. We have extracted 130 different features in time and frequency domain and selected the most efficient features with the sequential forward selection algorithm. With two stages of classification, the first is utilized for combining similar classes, and the second determines the final decision. We have tested two classifiers for our learning approach, the Gaussian mixture models (GMMs) and the hidden Markov models (HMMs) and compared their performances. We have improved the GMM classifier from our previous study and it has shown more promising results for this dataset. We achieved an accuracy of 88.92% when classifying the 8 unique activities with GMM and 93.5% with HMM when classifying 7 activities.

UR - http://www.scopus.com/inward/record.url?scp=85015411883&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015411883&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2016.64

DO - 10.1109/ICMLA.2016.64

M3 - Conference contribution

SP - 339

EP - 344

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

PB - Institute of Electrical and Electronics Engineers Inc.

ER -