Please use this identifier to cite or link to this item: https://hdl.handle.net/20.500.12313/1646
Title: A Personal Activity Recognition System Based on Smart Devices
Authors: Murcia, Harold F.
Triana, Juanita
Keywords: Activity recognition
machine learning
wearable devices
cell phone data
Myo armband
Issue Date: 16-Oct-2019
Publisher: Communications in Computer and Information Science
Citation: Murcia H., Triana J. (2019) A Personal Activity Recognition System Based on Smart Devices. In: Figueroa-García J., Duarte-González M., Jaramillo-Isaza S., Orjuela-Cañon A., Díaz-Gutierrez Y. (eds) Applied Computer Sciences in Engineering. WEA 2019. Communications in Computer and Information Science, vol 1052. Springer, Cham
Abstract: With the continuous evolution of technology, mobile devices are becoming more and more important in people’s lives. In the same way, new needs related to the information provided by their users arise, making evident the need to develop systems that take advantage of their daily use. The recognition of personal activity based on the information provided by the last generation mobile devices can easily be considered as an useful tool for many purposes and future applications. This paper presents the use of information provided from two smart devices in different acquisition schemes, assessing conventional supervised classifiers to recognize personal activity by an identification of seven classes. The classifiers were trained with a generated database from eight users and were evaluated in offline mode with other two generated databases. The prediction experiments were qualified by using F1-score indicator and were compared with the native prediction from the cellphone. The obtained results presented a maximum F1-score of 100% for the first validation test and 80.7% for the second validation test.
URI: https://hdl.handle.net/20.500.12313/1646
ISSN: 1865-0929
Appears in Collections:Artículos

Files in This Item:
There are no files associated with this item.


Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.