Using machine learning techniques to track individuals & their fitness activities
Reichherzer, T., Timm, M., Earley, N., Reyes, N., & Kumar, V. (2017). Using machine learning techniques to track individuals & their fitness activities. In A. Bossard, G. Lee, & L. Miller (Eds.), Proceedings of 32nd International Conference on Computers and Their Applications, March, 20-22, 2017, Honolulu, Hawaii, USA (pp. 119–124). Winona, MN, USA: ISCA.
Permanent Research Commons link: https://hdl.handle.net/10289/11113
The use of wearable devices for fitness and health tracking is on an upward curve with a range of devices now available from a number of manufacturers. The devices work with smart devices to exchange data via Bluetooth communication protocol. This paper presents the results of an initial study on the security and privacy weaknesses of wearable fitness devices. It discusses methods to 1) capture and process data sent from a wearable device to its paired smartphone during synchronization and 2) analyze the records to track individuals and make predictions. The data analysis methods use supervised machine-learning techniques to train a classifier for associating synchronization records with the individuals, their physical activities, and conditions under which they were performed. Results of the study show that the methods allow individuals and their activities to be tracked, both of which infringe on the privacy of the user. The paper also provides recommendations on improving the security of wearable devices based on the initial research results.
This is the author's accepted version. Copyright © 2017 by the International Society for Computers and Their Applications.