Application of a Cluster-Based Classifier Ensemble to Activity Recognition in Smart Homes

Anna Jurek, Yaxin Bi, Chris Nugent, S Wu

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

3 Citations (Scopus)

Abstract

An increasingly popular technique of monitoring activities within a smart environment involves the use of sensor technologies. With such an approach complex constructs of data are generated which subsequently require the use of activity recognition techniques to infer the underlying activity. The assignment of sensor data to one from a possible set of predefined activities can essentially be considered as a classification task. In this study, we propose the application of a cluster-based classifier ensemble method to the activity recognition problem, as an alternative to single classification models. Experimental evalua-tion has been conducted on publicly available sensor data collected over a period of 26 days from a single person apartment. Two types of sensor data representation have been considered, namely numeric and binary. The results show that the ensemble method can predict a numeric and binary representative activity with accuracies of 94.2% and 97.5%, respectively. These results outper-formed a range of single classifiers.
LanguageEnglish
Title of host publicationUnknown Host Publication
Pages88-95
Number of pages8
Publication statusPublished - Dec 2013
EventAmbient Assisted Living and Active Aging - 5th International Work-Conference, IWAAL 2013, Carrillo, Costa Rica, December 2-6, 2013 -
Duration: 1 Dec 2013 → …

Conference

ConferenceAmbient Assisted Living and Active Aging - 5th International Work-Conference, IWAAL 2013, Carrillo, Costa Rica, December 2-6, 2013
Period1/12/13 → …

Fingerprint

Classifiers
Sensors
Monitoring

Keywords

  • Activity recognition
  • classifier ensembles
  • smart homes

Cite this

@inproceedings{6b9108214db140568643c5c7fdadc134,
title = "Application of a Cluster-Based Classifier Ensemble to Activity Recognition in Smart Homes",
abstract = "An increasingly popular technique of monitoring activities within a smart environment involves the use of sensor technologies. With such an approach complex constructs of data are generated which subsequently require the use of activity recognition techniques to infer the underlying activity. The assignment of sensor data to one from a possible set of predefined activities can essentially be considered as a classification task. In this study, we propose the application of a cluster-based classifier ensemble method to the activity recognition problem, as an alternative to single classification models. Experimental evalua-tion has been conducted on publicly available sensor data collected over a period of 26 days from a single person apartment. Two types of sensor data representation have been considered, namely numeric and binary. The results show that the ensemble method can predict a numeric and binary representative activity with accuracies of 94.2{\%} and 97.5{\%}, respectively. These results outper-formed a range of single classifiers.",
keywords = "Activity recognition, classifier ensembles, smart homes",
author = "Anna Jurek and Yaxin Bi and Chris Nugent and S Wu",
year = "2013",
month = "12",
language = "English",
pages = "88--95",
booktitle = "Unknown Host Publication",

}

Jurek, A, Bi, Y, Nugent, C & Wu, S 2013, Application of a Cluster-Based Classifier Ensemble to Activity Recognition in Smart Homes. in Unknown Host Publication. pp. 88-95, Ambient Assisted Living and Active Aging - 5th International Work-Conference, IWAAL 2013, Carrillo, Costa Rica, December 2-6, 2013, 1/12/13.

Application of a Cluster-Based Classifier Ensemble to Activity Recognition in Smart Homes. / Jurek, Anna; Bi, Yaxin; Nugent, Chris; Wu, S.

Unknown Host Publication. 2013. p. 88-95.

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

TY - GEN

T1 - Application of a Cluster-Based Classifier Ensemble to Activity Recognition in Smart Homes

AU - Jurek, Anna

AU - Bi, Yaxin

AU - Nugent, Chris

AU - Wu, S

PY - 2013/12

Y1 - 2013/12

N2 - An increasingly popular technique of monitoring activities within a smart environment involves the use of sensor technologies. With such an approach complex constructs of data are generated which subsequently require the use of activity recognition techniques to infer the underlying activity. The assignment of sensor data to one from a possible set of predefined activities can essentially be considered as a classification task. In this study, we propose the application of a cluster-based classifier ensemble method to the activity recognition problem, as an alternative to single classification models. Experimental evalua-tion has been conducted on publicly available sensor data collected over a period of 26 days from a single person apartment. Two types of sensor data representation have been considered, namely numeric and binary. The results show that the ensemble method can predict a numeric and binary representative activity with accuracies of 94.2% and 97.5%, respectively. These results outper-formed a range of single classifiers.

AB - An increasingly popular technique of monitoring activities within a smart environment involves the use of sensor technologies. With such an approach complex constructs of data are generated which subsequently require the use of activity recognition techniques to infer the underlying activity. The assignment of sensor data to one from a possible set of predefined activities can essentially be considered as a classification task. In this study, we propose the application of a cluster-based classifier ensemble method to the activity recognition problem, as an alternative to single classification models. Experimental evalua-tion has been conducted on publicly available sensor data collected over a period of 26 days from a single person apartment. Two types of sensor data representation have been considered, namely numeric and binary. The results show that the ensemble method can predict a numeric and binary representative activity with accuracies of 94.2% and 97.5%, respectively. These results outper-formed a range of single classifiers.

KW - Activity recognition

KW - classifier ensembles

KW - smart homes

M3 - Conference contribution

SP - 88

EP - 95

BT - Unknown Host Publication

ER -