Experiement Data for Federated Learning

The datasets contain samples of accelerometer data from 13 volunteers that were used to demonstrate Federated learning approach. Datasets from 12 volunteers were used for the algorithm training, while a dataset from one participant was used for testing. The participants were asked to put the mobile phone running our data collector app in their pocket and perform four activities: sitting, walking, jogging, and standing. However, five of the participant could perform only three activities; Sitting, Standing and Walking. The participants whose dataset was used for testing performed the four activities. The mobile app collects the Ax, Ay, and Az axis data along with the timestamps at a frequency of 50Hz.