Train a deep neural network to classify human activity from sensor data recorded by a Samsung Galaxy SII.
Simply run 'predict_human_activity_boilerplate.py' and behold the 5 epochs of training! It'll tell you the accuracy score afterwards, it should be around 92%, but given the random nature of training, your mileage may vary.
The data is contained in the folder UCI_HAR_Dataset. It comes from the UC Irvine Machine Learning Repository at this website:
https://archive.ics.uci.edu/ml/datasets/human+activity+recognition+using+smartphones#
We are grateful to the creators of this dataset and the authors of the associated publications for making it freely available: Jorge L. Reyes-Ortiz, Davide Anguita, Alessandro Ghio, Luca Oneto and Xavier Parra.