Subject-dependent and -independent human activity recognition with person-specific and -independent models
Loading...
Files
Accepted version
Date
2019-09-16
Authors
Scheurer, Sebastian
Tedesco, Salvatore
Brown, Kenneth N.
O'Flynn, Brendan
Journal Title
Journal ISSN
Volume Title
Publisher
Association for Computing Machinery, ACM
Published Version
Abstract
The distinction between subject-dependent and subject-independent performance is ubiquitous in the Human Activity Recognition (HAR) literature. We test the hypotheses that HAR models achieve better subject-dependent performance than subject-independent performance, that a model trained with many users will achieve better subject-independent performance than one trained with a single user, and that one trained with a single user performs better for that user than one trained with this and other users by comparing four algorithms' subject-dependent and -independent performance across eight data sets using three different approaches, which we term person-independent models (PIMs), person-specific models (PSMs), and ensembles of PSMs (EPSMs). Our analysis shows that PSMs outperform PIMs by 3.5% for known users, PIMs outperform PSMs by 13.9% and ensembles of PSMs by a not significant 2.1% for unknown users, and that the performance for known users is 20.5% to 48% better than for unknown users.
Description
Keywords
Bagging , Boosting , Inertial sensors , Machine learning , Ensemble methods , Human activity recognition
Citation
Scheurer, S., Tedesco, S., Brown, K. N. and O'Flynn, B. (2019) Subject-dependent and -independent human activity recognition with person-specific and -independent models Proceedings of the 6th international Workshop on Sensor-based Activity Recognition and Interaction Rostock, Germany, 16-19 September. doi: 10.1145/3361684.3361689
Copyright
© 2019 Copyright held by the owner/author(s). Publication rights licensed to ACM. This is the author's version of the work. It is posted here for your personal use. Not for redistribution. The definitive Version of Record was published in HTTF 2019: Proceedings of the Halfway to the Future Symposium 2019, https://doi.org/10.1145/3363384.3363485